# Sleuth help center: every page The complete Sleuth help center in one file, updated November 9, 2026. The index is at llms.txt. --- # Welcome to Sleuth What Sleuth does, where you use it, and where to go in these docs for each job. [Start free](https://app.sleuth.studio/signup) Sleuth is AI software for biopharma strategy, business development, and competitive intelligence teams. You ask a question, like which indication to lead with or which pharma to pitch, and Sleuth builds the analysis to answer it: a dataset made for your question, with a source and a confidence score on every value, and a deliverable on top. ## What every plan includes Every plan, including the free trial, includes the same three parts: | Part | What you do with it | Uses credits? | | --- | --- | --- | | **Analysis built for your question** | Ask in your own words or start from a workflow. Sleuth plans the work, builds the datasets, and hands back an interactive analysis, a deck, or a spreadsheet. | Yes, one credit per analytical cell | | **An explorable knowledge base** | Browse and filter drugs, trials, deals, companies, and targets, the same records Sleuth works from, and build datasets from them. | No | | **Searchable documents** | Search and read sell-side research, filings, press releases, conference abstracts, and more. | No | ## Where you use it - **Sleuth Studio** is the web app at [app.sleuth.studio](https://app.sleuth.studio). Start here if you want to see the plan, the datasets, and the sources as Sleuth builds them. - **Claude, ChatGPT, Copilot, and other AI tools** connect to Sleuth through its MCP server. Ask from the tool you already use, and every project Sleuth runs is saved to your Studio account. See [Sleuth in AI tools](ai-tools-overview.md). - **Sleuth's analysts** can do the work for you with Concierge, or build a dataset to your spec and keep it current. See [Concierge and custom datasets](concierge-custom-datasets.md). ## What lives where | If you want to | Go to | | --- | --- | | Get set up and run something useful today | [Run your first analysis](first-analysis.md) | | Start from a common job, like choosing an indication | [Workflows](workflows-overview.md) | | Look up programs, trials, or deals yourself | [Explore the knowledge base](explore.md) | | Check what Sleuth built before you share it | [Check a dataset with Review mode](review-mode.md) | | Understand what you'll pay | [How credits work](how-credits-work.md) and [Plans and pricing](plans.md) | | Use Sleuth from Claude, ChatGPT, or Copilot | [Sleuth in AI tools](ai-tools-overview.md) | | Answer your security team's questions | [Security overview](security.md) | | Build on the MCP server | [MCP reference](mcp-overview.md) | ## Your first steps 1. **Sign up.** Start free at [app.sleuth.studio/signup](https://app.sleuth.studio/signup). The 14-day trial includes 2,500 credits and everything in Starter, and it doesn't need a card. [How sign-up works](sign-up.md) 2. **Tell Sleuth what you work on.** Onboarding asks about your role and the jobs you want done, so Sleuth can recommend workflows and remember your context. 3. **Run a workflow or ask a question.** Sleuth shows you its plan before it uses credits. [Walk through it](first-analysis.md) ## Related - [Run your first analysis](first-analysis.md): From sign-in to a finished analysis you can check and export, in five steps. - [How credits work](how-credits-work.md): What a credit buys, what is free, and how to tell what a job will use before it runs. - [Sleuth in Claude, ChatGPT, and Copilot](ai-tools-overview.md): Ask in the AI tool you already use, and Sleuth does the work behind the answer. *Updated November 9, 2026.* Source: md/welcome.md --- # Run your first analysis From sign-in to a finished analysis you can check and export, in five steps. [Open Sleuth Studio](https://app.sleuth.studio) This walks through one project start to finish in Sleuth Studio. If you'd rather ask from Claude, ChatGPT, or Copilot, set that up first with [Sleuth in AI tools](ai-tools-overview.md), then come back for steps 3 to 5. ## Before you start Pick a real question you'd otherwise spend days on. Sleuth works best on a decision with a clear scope: a modality and a target, an indication, a list of companies. "Which indications should our anti-CD19 cell therapy go after next?" works better than "tell me about autoimmune disease." ## 1. Choose how to start Home suggests workflows based on what you told Sleuth during onboarding, with the full library a click away. You can start in one of three places: - **A workflow,** if your job matches one, such as indication selection or asset assessment. You answer a few questions and Sleuth runs a plan built for that job. This is the fastest way to a useful result. [About workflows](workflows-overview.md) - **Analyze,** to ask anything in your own words. Each question becomes a project. [Ask a question](ask-question.md) - **Explore,** to filter the knowledge base yourself. Building datasets this way uses no credits. [Explore](explore.md) ## 2. Give Sleuth your context Say what you're working on, the decision you face, and anything you already know, like diseases you're weighing or criteria that matter to your team. Add a deck or a document if you have one. Sleuth reads it alongside its own sources and remembers it in [Your context](your-context.md), so you don't repeat yourself next time. ## 3. Approve the plan For anything bigger than a quick answer, Sleuth proposes a plan: the steps it will run and the datasets it will build, with the columns each one has. Read it, change what's wrong, and approve it. Nothing that uses credits runs before you approve, unless you've turned on auto-approve. In a workflow, choose **Pause at each step** to check the results as you go, or **Run end to end**, then press **Start**. [Plans and approvals](plans-approvals.md) ## 4. Watch it build Sleuth shows its reasoning in the chat as it works, then fills the datasets row by row. Larger landscapes take longer than a chat answer, because Sleuth reads filings, abstracts, and reports and checks each value. You can leave the page; Sleuth emails you when the project is done. ## 5. Check it, then take it with you 1. **Click any value** to see its source, the passage it came from, and a confidence score. 2. **Open Review mode** to see only the cells worth a second look, and mark each one reviewed or switch it to another source. [Review mode](review-mode.md) 3. **Export** the dataset to Excel, ask for slides in the chat and download the deck, or export the whole project for Claude or your team. [Outputs and exports](outputs.md) > **Good to know:** Your trial includes 2,500 credits. One credit is one analytical cell, like a judgment on whether one asset fits your strategy. Facts from the knowledge base, like phase and sponsor, are free. [How credits work](how-credits-work.md) ## Related - [Plans and approvals](plans-approvals.md): Sleuth shows you what it will build before it builds it. Here is what you approve and how to skip approvals. - [Check a dataset with Review mode](review-mode.md): See only the cells worth a second look, check their sources, and mark them reviewed. - [Workflows](workflows-overview.md): Six guided analyses for the decisions biopharma teams make most, each run on your own program, brief, or asset in four steps. *Updated November 9, 2026.* Source: md/first-analysis.md --- # Sign up and get access How to request an account, what approval looks like, and what onboarding asks. [Sign up](https://app.sleuth.studio/signup) ## Request access 1. **Open the sign-up page.** Go to [app.sleuth.studio/signup](https://app.sleuth.studio/signup), or press **Start free** anywhere on sleuthinsights.com. Enter your email and press **Continue**. 2. **Enter the code.** Sleuth emails you a 6-digit code that works for 10 minutes. If it doesn't arrive, check spam, then resend it; you can resend after 30 seconds, up to five times an hour. 3. **Tell us where you work.** Fill in your name, company, and country. With a work email, Sleuth asks for your company's website. With a personal email, like Gmail or Outlook, it asks for your LinkedIn profile. Press **Request access**. 4. **Wait for the email.** You'll see **You're on the list**. Sleuth's team approves new accounts by hand, and you'll get an email that says you're in. Sign in from its link. If your company already uses Sleuth, you may skip the wait: the last button reads **Create account** instead of **Request access**, and you go straight to onboarding. ## Onboarding Onboarding takes a few minutes and has six steps. Anything you entered at sign-up is filled in already, and you can skip the optional steps. | Step | What it asks | | --- | --- | | 1. Your workspace | Your workspace is set up for you. If colleagues from your company already have a Team workspace, you're offered to join it. | | 2. Your role | Your title and a line on what you do. | | 3. How you work | A few questions about how you like to work, which shape how Sleuth presents results. | | 4. What you want to get done | The jobs you bring to Sleuth, in priority order, like finding partners or benchmarking a program. | | 5. Where you'll use Sleuth | Studio, an AI tool like Claude, or both. Choose an AI tool and Sleuth walks you through connecting it. | | 6. Your team | On a Team plan, invite colleagues. On the trial and Starter, which are for one person, you'll see **Inviting teammates is on Team** and can press **Finish**. | Your answers are saved in [Your context](your-context.md), where you can change them later. When you finish, Home suggests workflows based on what you told it. ## Things to know - **Personal email addresses work.** Sign up with Gmail or Outlook and Sleuth asks for your LinkedIn profile instead of a company website. - **One trial per person.** Each email address gets one free trial. Sleuth treats name+tag addresses and dotted Gmail addresses as the same inbox. - **An email already on a Sleuth account can't sign up again.** Sign in instead, or ask your admin to invite you. - **You stay signed in for 30 days,** then you sign in again. - **Sign-up through an AI tool works too.** Connect Sleuth in Claude or another AI tool, and it asks for the same details in the chat, confirms each one with you, and sends your request. Two steps still happen in the browser, accepting the terms and setting up your workspace; then reconnect your AI tool. [Sleuth in AI tools](ai-tools-overview.md) ## Related - [Your free trial](free-trial.md): 14 days and 2,500 credits with everything in Starter, no card needed. - [Run your first analysis](first-analysis.md): From sign-in to a finished analysis you can check and export, in five steps. - [Invite your team](invite-team.md): Bring colleagues into a Team workspace so everyone draws on one pool of credits, fix an invite that didn't arrive, and remove someone who left. *Updated November 9, 2026.* Source: md/sign-up.md --- # Your free trial 14 days and 2,500 credits with everything in Starter, no card needed. [Start your trial](https://app.sleuth.studio/signup) ## What's included - **14 days and 2,500 credits.** Use them on anything: workflows, questions in Analyze, custom columns, and Sleuth in your AI tools. - **Everything in Starter.** The knowledge base, document search, sell-side research, monitoring, exports, and the MCP connection are all open. - **No card.** You only enter payment details if you choose a plan. - **One person.** The trial is a personal workspace, so you can't invite teammates. To try Sleuth as a team, choose a Team plan or [talk to us](https://www.sleuthinsights.com/talk-to-us) about a team trial. ## When it starts The 14 days start when your account is approved and your workspace is set up, not when you first request access. ## Keeping track The card at the bottom of the sidebar shows **Free trial**, how many credits you have left, and how many days. Project and dataset headers show your remaining credits too. Your balance can take a few hours to catch up with what you've used. ## If you run out of credits When your trial credits run out, starting a workflow, an export, or a chat message opens **You've used your trial credits**, with **See plans** and **Talk to us**. Anything already running finishes, and what you typed in the chat box stays there. Nothing is lost. Choose a plan and you can carry on right away. ## When the trial ends On day 15 your workspace becomes view only: - Your projects and datasets stay, and you can open and read them. They show a **View only** tag, and Studio says **Your free trial has ended. Your projects are still here to view.** - Chat, new runs, monitoring, downloads, and exports pause until you choose a plan. - Explore stays open. - A banner shows how many unused credits carry over. Choose a plan, during or after the trial, and everything works again. Trial credits you haven't used come with you and never expire. [Buy a plan](buy-plan.md) ## Raised under $10 million? Sleuth for Startups gives early-stage companies a free year for the whole team. [Sleuth for Startups](startups.md) ## Related - [Plans and pricing](plans.md): Starter for one person, Team for any number of people, Enterprise for company-wide use, and what a credit costs. - [Buy a plan](buy-plan.md): Choose a plan, pay by card, and what you get when the payment goes through. - [If you run out of credits](out-of-credits.md): What stops, what keeps working, and your options to carry on. *Updated November 9, 2026.* Source: md/free-trial.md --- # Find your way around Studio What each part of the Studio sidebar is for, and where settings, billing, and usage live. [Open Sleuth Studio](https://app.sleuth.studio) ## The sidebar | Item | What it's for | | --- | --- | | **Home** | See the workflows suggested for you, start something new, and pick up recent projects. | | **Workflows** | The full workflow library, grouped by the job you're doing. [Workflows](workflows-overview.md) | | **Explore** | Filter the knowledge base and build datasets yourself, without credits. [Explore](explore.md) | | **Analyze** | Ask anything. Each question becomes a project with its own chat, datasets, and results. [Ask a question](ask-question.md) | | **Document search** | Search and read the document library, sell-side research included. [Search documents](document-search.md) | | **Your context** | What Sleuth knows about you and your company, from onboarding and your uploads. [Your context](your-context.md) | | **Alerts** | What monitoring found in the projects you're tracking. [Monitoring](monitoring.md) | | **Usage and credits** | Your balance and what used it. [Balance and usage](balance-usage.md) | | **See it work** | Short videos of each workflow. [Videos](videos.md) | | **Connect via MCP** | Set up Sleuth in Claude, ChatGPT, Copilot, or another AI tool. [Sleuth in AI tools](ai-tools-overview.md) | | **Docs** | These docs. | | **Projects** | Your recent projects, with a count of new alerts on each. | ## The credit card in the sidebar At the bottom of the sidebar, a card shows your plan, the credits you have left, and when the next credits arrive. On a trial it also shows the days left. Admins see an **Upgrade** link there. ## Settings - **Billing** (admins): your plan, card, invoices, and plan changes. [Change your plan](change-plan.md) - **Usage** (admins): credits used by event, project, and person, with an Excel export. [Usage page](usage-page.md) - **Members** (admins on Team): who's in the workspace, and invites. [Invite your team](invite-team.md) - **Notifications**: which emails Sleuth sends you. [Notifications](notifications.md) - **Auto-approve**: let plans, dataset designs, and dataset changes go ahead without a click. [Plans and approvals](plans-approvals.md) ## On your phone Studio works on phones and tablets, so you can check a project or read results away from your desk. There's no app to install: Studio runs in your browser at [app.sleuth.studio](https://app.sleuth.studio). ## Related - [Run your first analysis](first-analysis.md): From sign-in to a finished analysis you can check and export, in five steps. - [Ask a question in Analyze](ask-question.md): How Sleuth turns a question in your own words into a plan, datasets, and an answer you can check. - [Explore the knowledge base](explore.md): Filter the records Sleuth works from and build your own dataset, with no credits. *Updated November 9, 2026.* Source: md/studio-tour.md --- # Videos Short clips that show Sleuth doing a real job, start to finish. [Book a walkthrough](https://www.sleuthinsights.com/talk-to-us) ## Start here - Video, coming soon: Sleuth from sign-in to export. One project, start to finish: a workflow, the plan, the dataset, checking it, and taking it into a meeting. ## One clip per workflow - Video, coming soon: Indication selection. Choose where a platform leads - Video, coming soon: Asset scouting. Score assets on your own brief - Video, coming soon: Pharma partnering. Line up pharma for a pitch - Video, coming soon: Asset assessment. Forecast an asset and set terms - Video, coming soon: Graveyard analysis. Why drugs like yours stopped - Video, coming soon: Standard of care evolution. Map the market you'll launch into ## Basics - Video, coming soon: Check a dataset in Review mode. Sources, scores, and the cells to check - Video, coming soon: Build a dataset in Explore. Filters, columns, and sending it to Analyze - Video, coming soon: Connect Sleuth to Claude. Set up, sign in, and ask your first question ## Rather see it live? Book 30 minutes with Sleuth's team to build your first project on your own question. [Talk to us](https://www.sleuthinsights.com/talk-to-us) ## Related - [Run your first analysis](first-analysis.md): From sign-in to a finished analysis you can check and export, in five steps. - [Workflows](workflows-overview.md): Six guided analyses for the decisions biopharma teams make most, each run on your own program, brief, or asset in four steps. - [Get help](get-help.md): Where to ask, depending on what you need. *Updated November 9, 2026.* Source: md/videos.md --- # Ask a question in Analyze How Sleuth turns a question in your own words into a plan, datasets, and an answer you can check. [Ask in Analyze](https://app.sleuth.studio) Analyze is where you work with Sleuth on a question. Each question becomes a project: the chat in the middle, the datasets beside it, and the results as Sleuth builds them. ## What happens when you ask Sleuth reads your message and decides what kind of request it is: | You ask | Sleuth | | --- | --- | | A quick question, like "What phase is zanidatamab in for biliary tract cancer?" | Answers in the chat, with sources. | | A question that needs a dataset, like "Rank every bispecific in Phase 2 or later for solid tumors on differentiation" | Proposes a plan and the datasets it will build, and waits for you to approve them. [Plans and approvals](plans-approvals.md) | | A question about data already in the project | Answers from your datasets, without rebuilding them. | | A change, like "Add a column for each asset's lead indication" | Proposes the change to the dataset and makes it once you approve. | If your question is unclear, Sleuth asks before it starts. Answer in the chat and it carries on. ## Write a question Sleuth can run with - **Name the decision.** "We need to pick two partners to approach at BIO" gives Sleuth more to work with than "find partners." - **Set the scope.** A modality, a target, an indication, a stage, a region, or a list of companies. - **Say what counts.** If unmet need matters more than market size to you, say so. Sleuth builds those criteria into the columns. - **Add what you have.** Upload a deck, a shortlist, or a spreadsheet, and Sleuth uses it. [Use your own documents and data](own-data.md) ## While it works Sleuth shows its reasoning as it goes, then fills each dataset. A project can hold several linked datasets, each at its own level of detail, such as one for programs and one for deals. You can leave; Sleuth emails you when it's done. The chat panel has a few tools to keep long projects manageable: pins for results you want to keep in view, a Files tab for your uploads and outputs, a task panel that shows what's running, and suggested follow-ups you can click. ## Keep going in the same project Ask follow-up questions, ask for a chart or a deck, add columns, or narrow the rows. Questions about results you already have don't use credits. If a dataset changes after an answer was built, the answer shows that it's based on older data, with a **Regenerate** button. Sleuth also remembers what you tell it within a project, and preferences you set across projects, so you don't have to repeat your scope or your criteria. ## Related - [Plans and approvals](plans-approvals.md): Sleuth shows you what it will build before it builds it. Here is what you approve and how to skip approvals. - [Datasets: rows, columns, and cells](datasets.md): What a Sleuth dataset is, which cells use credits, and how to change one. - [Outputs and exports](outputs.md): What Sleuth hands back, and how to get each one out of Sleuth. *Updated November 9, 2026.* Source: md/ask-question.md --- # Plans and approvals Sleuth shows you what it will build before it builds it. Here is what you approve and how to skip approvals. [Open Analyze](https://app.sleuth.studio) ## What you approve A bigger question in Analyze pauses twice before it uses credits: 1. **The plan.** The steps Sleuth will run, in order, and the datasets it will build. Approve it, or reject it and say what to change, like "skip deals, focus on Phase 2 and later." 2. **The dataset design.** The columns each dataset will have and what goes in each one. This is where you add a criterion, drop a column you don't need, or rename one to match how your team talks. If you don't approve, nothing runs and nothing is charged: the plan waits in the project until you approve it or ask for a change. After you approve both, Sleuth builds the datasets and the analysis. Later, when you ask Sleuth to change a dataset, it shows you the change and waits for your OK. ## Turn on auto-approve If you'd rather Sleuth go straight through, turn on auto-approve in Settings. You can choose it separately for plans, dataset designs, and dataset changes. Runs you start from Claude, ChatGPT, or Copilot follow the same settings, so with auto-approve off they wait for your approval there too. > **Good to know:** Approving the plan is the best moment to check scope. A plan that covers "every ADC in solid tumors" when you meant HER2-low breast cancer builds a much bigger dataset and uses more credits. ## Workflows have one plan screen Workflows show the whole plan on one screen before they run: the data steps, the outputs, and your credit balance. Add or remove outputs, then choose how it runs: - **Pause at each step** (recommended): Sleuth stops after each step so you can check the results, pick an option, or skip ahead. - **Run end to end**: Sleuth takes the defaults, lists them in the chat, and runs straight through. Nothing runs until you press **Start**. [How workflows run](workflows-overview.md) ## Related - [Ask a question in Analyze](ask-question.md): How Sleuth turns a question in your own words into a plan, datasets, and an answer you can check. - [How credits work](how-credits-work.md): What a credit buys, what is free, and how to tell what a job will use before it runs. - [How a research run works](mcp-run-lifecycle.md): Create a project, send the question, wait for the status to change, handle approvals and clarifications, then read the datasets and the answer. *Updated November 9, 2026.* Source: md/plans-approvals.md --- # Datasets: rows, columns, and cells What a Sleuth dataset is, which cells use credits, and how to change one. [Open Analyze](https://app.sleuth.studio) Every analysis in Sleuth sits on a dataset: a table where each row is a record, like a drug program, a company, or a trial, and each column answers part of your question. ## Two kinds of cells | Kind | Example | Credits | | --- | --- | --- | | **Fact cells**: values Sleuth's knowledge base already has | An asset's sponsor, target, modality, and highest phase | Free | | **Analytical cells**: judgments Sleuth makes for your question | How well an asset fits your strategy, or whether a program has white space in 2L NSCLC | 1 credit each | So scoring 150 assets on 3 criteria uses 450 credits, while their phase and sponsor cost nothing. [How credits work](how-credits-work.md) ## What's in every cell Click any cell to see its value, the source it came from, the passage behind it, and a confidence score. Choose **See more** for the score breakdown and Sleuth's explanation. [Confidence scores](confidence.md) ## Change a dataset - **Edit a value** by double-clicking the cell. Your correction gets a confidence of 100. - **Re-run a cell** with new guidance, like "use the label, not the press release." - **Add rows or columns** by asking in the chat, or with **Add**. New AI-filled columns are custom columns. [Add your own columns](custom-columns.md) - **Delete** rows or columns you don't need. You can't edit a past version of a dataset, or a dataset someone else owns. ## Versions When monitoring or an edit changes a dataset, Sleuth saves a new version with a log of what changed. You can open any earlier version. [Monitoring](monitoring.md) ## Download it Download any dataset as a CSV that opens in Excel. [Outputs and exports](outputs.md) ## Related - [Add your own columns](custom-columns.md): Ask Sleuth to fill in a column you define, on any dataset in Explore or Analyze. - [Confidence scores and labels](confidence.md): How Sleuth scores every value, and what sourced, inferred, and unknown mean. - [Check a dataset with Review mode](review-mode.md): See only the cells worth a second look, check their sources, and mark them reviewed. *Updated November 9, 2026.* Source: md/datasets.md --- # Add your own columns Ask Sleuth to fill in a column you define, on any dataset in Explore or Analyze. [Open Explore](https://app.sleuth.studio) A custom column is a question you ask of every row. "Does this program have human efficacy data?" or "Which pharma licensed this asset, if any?" Sleuth fills it in for each row, with a value, its reasoning, a confidence score, and the sources. ## Add a column 1. **Open the dataset** in Explore or Analyze and choose **Add**, then columns. 2. **Describe the column.** Give it a name and say what it should contain, with any rules, like "Yes only if a Phase 2 readout reported a response rate." 3. **Pick the type** of answer: text, a number, yes or no, one choice from a list, or several. 4. **Run it.** Sleuth shows progress as it fills each row. You can add rows the same way, by describing what to add. ## Credits Each cell Sleuth fills in a custom column is an analytical cell and uses one credit. A column across 200 rows uses 200 credits. ## Use it in the analysis After you add columns, re-run the analysis so it uses them. The charts and results update without rebuilding the rest of the project. ## Related - [Datasets: rows, columns, and cells](datasets.md): What a Sleuth dataset is, which cells use credits, and how to change one. - [Explore the knowledge base](explore.md): Filter the records Sleuth works from and build your own dataset, with no credits. - [How credits work](how-credits-work.md): What a credit buys, what is free, and how to tell what a job will use before it runs. *Updated November 9, 2026.* Source: md/custom-columns.md --- # Outputs and exports What Sleuth hands back, and how to get each one out of Sleuth. [Open a project](https://app.sleuth.studio) Every output comes from the same datasets, so a number in the deck matches the number in the spreadsheet, and both link to their sources. ## What you can get | Output | What it is | How to get it | | --- | --- | --- | | **Interactive analysis** | A view you click through: a landscape, a ranking, a map. Click a company or program to see its row and sources. | Built in the project. Ask for a different view in the chat. | | **Deck** | PowerPoint slides for the meeting, built from the same rows. | Ask for slides, then download from the project's Files tab. | | **Report** | A written analysis as a web page you can share. | Ask for a report, then open or download it. | | **Charts** | Each chart, with the source behind every data point. | Export as PNG or PDF. | | **Spreadsheet** | Every row and column, plus the working data behind the analysis. | Download any dataset or table as a CSV. | | **Project export** | One ZIP with the whole project: your questions, context, datasets, and outputs, plus a file for each dataset listing every cell's value, confidence, and sources. Up to 10,000 rows per dataset. | Choose export on the project. Hand it to Claude, ChatGPT, your team, or anyone who needs to see where each value came from. | ## Use your own slide template Download the dataset and the chart together, then rebuild the chart in your template. The CSV has every value the chart uses. ## Hand a project to another AI tool Project export packs everything an AI tool needs to keep working on your project. If you use Sleuth in Claude or another AI tool, you don't need the export: the AI tool can read your projects directly. [Sleuth in AI tools](ai-tools-overview.md) ## Related - [Datasets: rows, columns, and cells](datasets.md): What a Sleuth dataset is, which cells use credits, and how to change one. - [Projects, initiatives, and sharing](share-work.md): How projects are organized, who can see them, and how to get work to people outside Sleuth. - [Sleuth in Claude, ChatGPT, and Copilot](ai-tools-overview.md): Ask in the AI tool you already use, and Sleuth does the work behind the answer. *Updated November 9, 2026.* Source: md/outputs.md --- # Keep a project current with monitoring Turn on monitoring and Sleuth updates your datasets as new trials, filings, and deals come in. [Open a project](https://app.sleuth.studio) ## Turn it on Switch on monitoring for a dataset, or for the project's results as a whole. Sleuth works out what matters from the project itself. You can add focus instructions, like "flag new Phase 3 starts and any licensing deal over $50M upfront." ## What happens next - **Sleuth watches the knowledge base** for new information about the rows you're tracking, and for new rows that fit your criteria. - **Your dataset updates.** Each update saves a new version with a log of what changed and where it came from. You can open any earlier version. - **Alerts arrive.** Each project gets an Alerts tab, and the sidebar counts new alerts. Important ones also come by email, depending on your notification settings. ## Credits Monitoring only uses credits when it finds something new and Sleuth analyzes it. A check that finds nothing costs nothing. A month of monitoring typically adds dozens to hundreds of credits, depending on how big the dataset is and how often it changes, so monitor what you'll act on and turn it off when a project is done. There's no limit on how many projects you monitor. ## Turn it off Switch monitoring off on the dataset or project. Your data and versions stay. > **Good to know:** If your workspace runs out of credits, or your trial or plan ends, monitoring pauses until credits are back. ## Related - [Datasets: rows, columns, and cells](datasets.md): What a Sleuth dataset is, which cells use credits, and how to change one. - [Notifications](notifications.md): The emails Sleuth sends and how to change them. - [How fresh the data is](data-freshness.md): How quickly new trials, filings, and deals reach Sleuth, and how to keep your own work current. *Updated November 9, 2026.* Source: md/monitoring.md --- # Use your own documents and data Upload files for Sleuth to read, start from your own spreadsheet, or have Sleuth source the values you already have. [Open Analyze](https://app.sleuth.studio) ## Upload documents Drag files into a project, or select several at once: PDFs, slide decks, Word documents, and Markdown. Sleuth reads them alongside its own sources and cites them like any other source. A pipeline deck, a target list, or a competitor's poster all work. ## Start from your own spreadsheet Upload a CSV or Excel table to start a project from it. Sleuth treats it like any dataset it built: you can add columns, add rows, and run an analysis on it. You can also start from a dataset you built in Explore. [Explore](explore.md) ## Source what you already have If you upload a table you've built by hand, ask Sleuth to verify it. Sleuth finds sources for the values already in the table without changing them, so you can see which ones hold up. Each cell it checks uses one credit. ## File sizes and what not to upload - From Claude and other AI tools, each file can be up to 10 MB. Upload bigger files in Studio. - Don't upload patient health information. Sleuth's terms don't allow it. ## Save it for next time Documents about your company, like your pipeline deck, belong in [Your context](your-context.md), where every workflow and chat can use them. When you upload a file to a workflow, Sleuth offers to save it there. ## Who can see your files Files you upload stay in your workspace and are never shown to other customers. [How your data is used](data-use.md) > **On Enterprise:** Sleuth can connect to Drive, SharePoint, Box, or your data warehouse, and match your internal program names to the same records. [Enterprise](enterprise.md) ## Related - [Your context](your-context.md): What Sleuth keeps about you and your company so you never have to explain it twice. - [How your data is used](data-use.md): Who can see your questions, uploads, and results, and how to delete them. - [Datasets: rows, columns, and cells](datasets.md): What a Sleuth dataset is, which cells use credits, and how to change one. *Updated November 9, 2026.* Source: md/own-data.md --- # Your context What Sleuth keeps about you and your company so you never have to explain it twice. [Open Your context](https://app.sleuth.studio) Your context is the part of Studio that holds what Sleuth knows about you and your company. Every workflow, every chat, and Sleuth in your AI tools read from it. ## What's in it - **Profile.** Your company, role, the jobs you bring to Sleuth, therapeutic areas, and modalities, from onboarding. Edit any of it. - **Documents you gave Sleuth.** Decks and documents about your company, each with what Sleuth took from it, so you can check it read them right. - **Learned from your projects.** Facts and preferences Sleuth picked up in your work, like "we only consider assets with US rights available." Sleuth saves nothing here without your OK: choose **Keep** or **Drop**. ## The best thing to add One good document about your company: your platform, your programs, what you've shown so far, and what you're deciding next. Sleuth flags documents that look out of date, like a deck older than three months, and asks for a newer one. ## Who can see it Your context is private to your workspace. The page shows who it's shared with, for example everyone in your Team workspace. ## Related - [Use your own documents and data](own-data.md): Upload files for Sleuth to read, start from your own spreadsheet, or have Sleuth source the values you already have. - [How your data is used](data-use.md): Who can see your questions, uploads, and results, and how to delete them. - [Sign up and get access](sign-up.md): How to request an account, what approval looks like, and what onboarding asks. *Updated November 9, 2026.* Source: md/your-context.md --- # Projects, initiatives, and sharing How projects are organized, who can see them, and how to get work to people outside Sleuth. [Open Sleuth Studio](https://app.sleuth.studio) ## Projects Every question, workflow run, or Explore dataset lives in a project, with its chat, datasets, files, alerts, and outputs. Projects you start from Claude, ChatGPT, or Copilot are saved to your account the same way, so you can open them in Studio later. ## Initiatives An initiative is a folder for related projects and their datasets, like everything for one asset or one deal. Use initiatives to keep a Team workspace tidy. ## Who can see a project Each project belongs to the person who made it, on every plan. On a Team plan, everyone draws on the same credits, and you choose which projects to share. ## Work on a project together | To | Use | What happens | | --- | --- | --- | | Give a colleague a copy to build on | **Clone project** | Your colleague gets their own copy, and you keep yours. | | Hand a project over | **Transfer project** | Your colleague owns it, and it leaves your projects. | ## Share with people outside Sleuth Download the deck, the report, the charts, or the dataset and send them on. To hand over everything, export the project as one ZIP, which lists every cell's sources. [Outputs and exports](outputs.md) ## Related - [Outputs and exports](outputs.md): What Sleuth hands back, and how to get each one out of Sleuth. - [Invite your team](invite-team.md): Bring colleagues into a Team workspace so everyone draws on one pool of credits, fix an invite that didn't arrive, and remove someone who left. - [How your data is used](data-use.md): Who can see your questions, uploads, and results, and how to delete them. *Updated November 9, 2026.* Source: md/share-work.md --- # Notifications The emails Sleuth sends and how to change them. [Open Studio settings](https://app.sleuth.studio) ## Emails Sleuth sends | Email | When | Who gets it | | --- | --- | --- | | Your dataset or analysis is ready | A build finishes. A project with several datasets sends one email listing each. | The person who started it | | Monitoring alert | Monitoring finds something important in a project you track | The project's owner | | Credit alerts | Your workspace has used 75% and 90% of its credits, and when they run out | The workspace admin | | You're in | Your access request is approved | You | | Receipts and invoices | Each payment | The person who pays | | Payment problem | A renewal charge fails | The person who pays | ## Change what you get Open Settings and choose which emails you want. Emails about billing and access always go out. ## Related - [Keep a project current with monitoring](monitoring.md): Turn on monitoring and Sleuth updates your datasets as new trials, filings, and deals come in. - [Check your balance and usage](balance-usage.md): Where to see your credits, how alerts work, and why the balance can lag. - [Invoices, cards, and payment problems](payment-problems.md): Change your card, find invoices, and what happens if a renewal payment fails. *Updated November 9, 2026.* Source: md/notifications.md --- # Explore the knowledge base Filter the records Sleuth works from and build your own dataset, with no credits. [Open Explore](https://app.sleuth.studio) Explore is for when you know the table you want. Pick filters, choose columns, and Sleuth builds a dataset from its knowledge base, with a source and a confidence score on every value. Browsing and building datasets in Explore never use credits. ## Build a dataset 1. **Start from drugs.** Each row is a drug program. More starting points, like trials, companies, and indications, are coming. 2. **Add filters.** Filter by target, modality, indication, therapeutic area, company, and phase of development. Indications are organized as a tree, so you can pick a whole disease area or one subtype. Search inside any filter by name or synonym. 3. **Choose columns.** Each starting point comes with sensible default columns. Show, hide, and sort them. The table previews live as you go. 4. **Generate it.** Sleuth builds the dataset and tells you if it hit the row limit, so you know to narrow your filters. Very broad filters, like every program in lung cancer, can take too long; a narrower indication or a single modality works better. ## Switch the view Look at the same dataset as assets, trials, or deals without rebuilding it. The trials view lists each program's studies; the deals view lists the licensing deals, acquisitions, and collaborations around them. ## Save and reuse - **Save a view** to come back to the same filters and columns, and duplicate it to try a variation. - **Filter within a column** once the dataset is built. - **Export** the dataset as a CSV. ## Take it further - **Add your own columns** for anything the knowledge base doesn't have, like a judgment on each asset. These use credits. [Add your own columns](custom-columns.md) - **Send it to Analyze** to ask questions of it, score it, or build a deck on it. The dataset links to every Analyze project made from it. > **In AI tools:** Claude and other AI tools can build Explore datasets too: describe the filters in plain words and the AI tool returns the counts, a preview, and a download link. [Explore tools](tools-explore.md) ## Related - [What's in the knowledge base](knowledge-base.md): The records Sleuth works from, how many there are, and how they connect. - [Add your own columns](custom-columns.md): Ask Sleuth to fill in a column you define, on any dataset in Explore or Analyze. - [Search documents](document-search.md): Search Sleuth's document library, read any result in full, and see exactly where a value came from. *Updated November 9, 2026.* Source: md/explore.md --- # Search documents Search Sleuth's document library, read any result in full, and see exactly where a value came from. [Open document search](https://app.sleuth.studio) Sleuth's knowledge base is built from more than 150 million documents. Document search lets you search them directly. Searching and reading never use credits. ## What you can search | Category | For example | | --- | --- | | Broker research | Sell-side initiations, earnings notes, conference takeaways, forecasts [More](sell-side.md) | | Company filings | 10-Ks, 8-Ks, licensing agreements filed as exhibits, Hong Kong and Asian exchange filings | | Company presentations and transcripts | Investor decks, R&D day slides, earnings call transcripts | | Scientific literature | Papers, preprints, and conference abstracts and posters | | Regulatory | Labels, approval and complete response letters, advisory committee materials, guidance | | Patents | US, European, Japanese, Chinese, and international applications | | News | Press releases and trade press | | Other documents | Everything else Sleuth reads | ## Build a search - **Keywords.** Add terms, and combine rows with AND and OR. "claudin 18.2 AND gastric" finds documents with both. - **Categories.** Limit the search to one or more categories, like broker research only. - **Dates.** Set a date range, or sort by date to see what's new. - **Recent uploads.** See the documents most recently added to a category. Sleuth sometimes adds older documents, so check each one's date. ## Read the result Results open inside Sleuth at the pages that match, so you can read the passage in context and the rest of the document. When a dataset cell cites a document, clicking the citation opens the same view. ## In AI tools Claude and other AI tools can search the same library, list the documents most recently added to a category, and pull the relevant passages from a document. [Research library tools](tools-research.md) ## Related - [Sell-side research](sell-side.md): Research from 35+ banks and brokers, included in every plan, searchable next to everything else. - [What's in the knowledge base](knowledge-base.md): The records Sleuth works from, how many there are, and how they connect. - [Every source](sources.md): The registries, regulators, journals, filings, and feeds Sleuth reads, grouped by what they answer. *Updated November 9, 2026.* Source: md/document-search.md --- # Sell-side research Research from 35+ banks and brokers, included in every plan, searchable next to everything else. [Search broker research](https://app.sleuth.studio) Every Sleuth plan includes sell-side research from 35+ providers. You don't need an account with each bank, and there's no extra charge. ## Who it's from | Provider type | Covers | | --- | --- | | Global investment banks | Large-cap pharma and biotech | | US healthcare and mid-market banks | Small- and mid-cap biotech | | European brokers | European pharma and biotech | | Independent and sponsored research | Small-cap and non-US-listed biotechs | | Chinese domestic brokers | Chinese pharma and biotech, in English | | Consulting and industry research | Industry reports and sector analysis | | Historical archives | Past research back to 2017 | That's over 500,000 reports, including research from China's domestic brokers, in English, and new notes arrive within 48 hours. For the full list of providers, [ask us](https://www.sleuthinsights.com/talk-to-us). ## Already get these reports by email? Your inbox holds what your own banks send you. Sleuth puts research from 35+ providers in one place, searchable alongside filings, trials, and news, and reads it when it builds your datasets, so a value like a peak sales estimate comes with the passage it came from. ## What kinds of reports Initiations and company deep dives, earnings previews and recaps, therapeutic area primers, conference takeaways, KOL calls and physician surveys, drug-level sales forecasts and consensus estimates, price targets and valuation models, and M&A and licensing commentary. ## How you use it - **Search and read** any report in full inside Sleuth, and jump to the passage that matches. - **Let Sleuth use it.** Sleuth reads the reports when it builds a dataset and cites the exact passage behind a value, like a peak sales estimate. - **Download what Sleuth built from it.** Datasets built from the reports download like any other dataset. The original report files stay in Sleuth: you can read every report, but you can't download the PDF. ## In AI tools When you search broker research from Claude or another AI tool, Sleuth returns the relevant passages and page images, not the whole PDF. ## Related - [Search documents](document-search.md): Search Sleuth's document library, read any result in full, and see exactly where a value came from. - [Every source](sources.md): The registries, regulators, journals, filings, and feeds Sleuth reads, grouped by what they answer. - [How your data is used](data-use.md): Who can see your questions, uploads, and results, and how to delete them. *Updated November 9, 2026.* Source: md/sell-side.md --- # What's in the knowledge base The records Sleuth works from, how many there are, and how they connect. [See the data page](https://www.sleuthinsights.com/data) Sleuth's knowledge base is all of Sleuth's data, connected: every record is matched across its sources and linked to the documents behind it. Studio, Sleuth in your AI tools, and Sleuth's analysts all work from the same knowledge base. ## The records | Record | Count | What each one has | | --- | --- | --- | | **Drugs** | 150K+ | Every program by target, modality, indication, and phase, with catalysts and sales forecasts. About 60% are preclinical, including programs that so far appear only in patents. | | **Trials** | 1M+ | From 20 registries, each linked to its drugs, sponsors, sites, and results. | | **Companies** | 400K+ | Public and private, with pipelines, deals, and financings, and both Chinese and English names for Chinese companies. | | **Targets** | 500K+ | Biological targets across every major therapeutic area, matched to one dictionary. | | **Deals** | 200K+ | Licensing, M&A, collaborations, and financings, with upfronts, milestones, and royalties where they're disclosed. | | **Documents** | 150M+ | The filings, papers, abstracts, reports, and releases behind every record. [Search them](document-search.md) | *Counts as of November 2026.* ## Where it comes from Sleuth collects from the organizations that publish the data: trial registries, regulators, patent offices, journals and medical meetings, companies themselves, news, and sell-side research, including China's own sources, read in Chinese. [See every source](sources.md) ## Why connected matters A drug shows up under many names: a code, an INN, a brand, a Chinese name. Sleuth matches them to one record, so a trial in China's registry, a deal in an 8-K, and a poster at ASCO all land on the same drug. That's what lets one question reach all of them. ## Does it replace a database subscription? For most strategy work, yes. Every plan includes the same kinds of records a database like Evaluate, Cortellis, Citeline, or GlobalData has, plus sell-side research and China coverage, and browsing is free. Then Sleuth does the narrowing down you'd otherwise do in Excel. ## Related - [Every source](sources.md): The registries, regulators, journals, filings, and feeds Sleuth reads, grouped by what they answer. - [How fresh the data is](data-freshness.md): How quickly new trials, filings, and deals reach Sleuth, and how to keep your own work current. - [Explore the knowledge base](explore.md): Filter the records Sleuth works from and build your own dataset, with no credits. *Updated November 9, 2026.* Source: md/knowledge-base.md --- # Every source The registries, regulators, journals, filings, and feeds Sleuth reads, grouped by what they answer. [See the data page](https://www.sleuthinsights.com/data) Sleuth collects every source below at the source and keeps it current. Each one feeds the records in the knowledge base. Use the filter to find a source by name or region. ### Clinical trials and registries Who is running what study, where, and at what stage. | Source | Region | What it covers | | --- | --- | --- | | **ClinicalTrials.gov** | Global | Trial registrations, status, arms, endpoints, sponsors, sites and posted results | | **WHO ICTRP** | Global | Aggregated records from WHO primary registries worldwide | | **EU Clinical Trials Register (EudraCT)** | EU | EU trials started before the Clinical Trials Regulation | | **EU CTIS** | EU | EU trial applications, decisions and results under the Clinical Trials Regulation | | **ISRCTN** | UK | UK primary registry | | **DRKS** | EU | German Clinical Trials Register | | **ChiCTR** | China | Chinese Clinical Trial Registry | | **China Drug Trial Registration Platform (CDE)** | China | Mandatory registration for drug trials in China, often earlier and more complete than ChiCTR for industry trials | | **ITMCTR** | China | International Traditional Medicine Clinical Trial Registry, via WHO ICTRP | | **jRCT** | Japan | Japan Registry of Clinical Trials | | **ANZCTR** | APAC | Australian New Zealand Clinical Trials Registry | | **CTRI** | APAC | Clinical Trials Registry – India | | **ReBEC** | LatAm | Brazilian Clinical Trials Registry, via WHO ICTRP | | **CRIS** | Korea | Korea's Clinical Research Information Service, via WHO ICTRP | | **IRCT** | Middle East and Africa | Iranian Registry of Clinical Trials, via WHO ICTRP | | **PACTR** | Middle East and Africa | Pan African Clinical Trials Registry, via WHO ICTRP | | **NL-OMON** | EU | Overview of Medical Research in the Netherlands, via WHO ICTRP | | **RPCEC** | LatAm | Cuban Public Registry of Clinical Trials, via WHO ICTRP | | **REPEC** | LatAm | Peruvian Clinical Trials Registry, via WHO ICTRP | | **SLCTR** | APAC | Sri Lanka Clinical Trials Registry, via WHO ICTRP | | **TCTR** | APAC | Thai Clinical Trials Registry, via WHO ICTRP | ### Scientific literature and preprints What the published science says. | Source | Region | What it covers | | --- | --- | --- | | **PubMed / MEDLINE** | Global | Biomedical citations and abstracts | | **PubMed Central Open Access** | Global | Full-text open-access articles | | **Select peer-reviewed clinical papers** | Global | Clinical trial publications beyond PubMed's open-access set | | **Europe PMC** | EU | European full text, with links to grants | | **bioRxiv** | Global | Biology preprints | | **medRxiv** | Global | Clinical and health-science preprints | | **Chinese-language journals, theses, and conference papers** | China | Chinese clinical and pharmacology research rarely indexed outside China | | **Crossref** | Global | DOI metadata across journals, books and proceedings | | **OpenAlex** | Global | Open index of papers, authors and institutions, used for KOL and lab mapping | ### Regulatory What was approved, designated, delayed, or rejected, by which agency. | Source | Region | What it covers | | --- | --- | --- | | **FDA drug labels (openFDA / SPL)** | US | Prescribing information, indications, boxed warnings | | **DailyMed** | US | Current structured product labels from the National Library of Medicine | | **Drugs@FDA** | US | Approval history, approval letters and review packages for NDAs and BLAs | | **FDA approval and decision letters** | US | Approval letters and agency communications | | **FDA Complete Response Letters** | US | Published CRLs explaining why applications were not approved | | **FDA Advisory Committee materials** | US | Briefing documents, transcripts, minutes and votes | | **Orange Book** | US | Approved small-molecule drugs with patents, exclusivities and therapeutic equivalence | | **Purple Book** | US | Licensed biologics, biosimilars, interchangeables and reference product exclusivity | | **CDER Novel Drug Approvals** | US | Annual lists of new molecular entities and new biologics | | **CBER approved biologics and cell & gene therapies** | US | Vaccines, blood products, cellular and gene therapy approvals | | **FDA Oncology approval notifications (OCE)** | US | Oncology and hematology approvals as they happen | | **FDA press announcements** | US | Agency news releases | | **Postmarketing Requirements and Commitments** | US | Required follow-on studies and their status | | **REMS@FDA** | US | Risk evaluation and mitigation strategies | | **FDA guidance documents** | US | Draft and final guidance on drug development and review | | **Product-specific guidances** | US | Bioequivalence guidance for generics | | **FDA notices in the Federal Register** | US | Approval withdrawals, advisory committee notices, exclusivity and user fee notices | | **Orphan Drug Designations and Approvals database** | US | Every orphan designation with sponsor, date and indication | | **Breakthrough Therapy designation lists (CDER, CBER)** | US | Breakthrough requests granted and resulting approvals | | **Accelerated Approval lists** | US | Accelerated approvals, confirmatory trial status and withdrawals | | **RMAT designations (CBER)** | US | Regenerative medicine advanced therapy designations | | **Rare Pediatric Disease designations** | US | Designations eligible for priority review vouchers | | **Priority review vouchers** | US | Vouchers granted, used and sold (rare pediatric, tropical disease, countermeasure) | | **EMA CHMP opinions and meeting highlights** | EU | Monthly positive and negative opinions | | **EMA European Public Assessment Reports (EPARs)** | EU | Full assessment reports for authorized medicines | | **EC Union Register of medicinal products** | EU | EU marketing authorizations and decisions | | **EU orphan designations (Community Register)** | EU | Orphan designations by sponsor and condition | | **EMA PRIME** | EU | Medicines granted priority medicines support | | **EMA Paediatric Investigation Plans** | EU | PIP decisions and waivers | | **EMA withdrawn applications** | EU | Applications pulled before an opinion | | **MHRA** | UK | UK approvals and product information | | **NMPA** | China | Drug and biologic approvals and announcements | | **CDE** | China | IND acceptances, priority review, breakthrough therapy and conditional approval lists; review reports | | **PMDA** | Japan | Approvals list and review reports | | **MHLW designations** | Japan | Sakigake, orphan and conditional approval designations | | **Health Canada Drug Product Database** | Canada | Drugs marketed in Canada | | **Health Canada Notice of Compliance database** | Canada | Canadian approvals, including conditional NOCs | | **TGA Australian Register of Therapeutic Goods (ARTG)** | APAC | Australian registrations | | **TGA Australian Public Assessment Reports (AusPARs)** | APAC | Public assessment reports for Australian approvals | | **Project Orbis** | Global | FDA-led concurrent review of cancer drugs with partner agencies | | **Access Consortium** | Global | Work-sharing reviews by the UK, Canada, Australia, Singapore, and Switzerland | ### Safety and pharmacovigilance What safety signals exist after approval. | Source | Region | What it covers | | --- | --- | --- | | **FAERS (via openFDA)** | US | Adverse event reports for drugs and therapeutic biologics | | **FDA Drug Safety-related Labeling Changes (SrLC)** | US | Label safety updates by drug and date | | **FDA Drug Safety Communications** | US | Agency safety notices by drug | | **MedWatch safety alerts** | US | FDA safety alerts for drugs and biologics | | **VAERS** | US | Vaccine adverse event reports | | **EudraVigilance (adrreports.eu)** | EU | Suspected adverse reaction reports in the EU | | **EMA PRAC signals and meeting highlights** | EU | Safety signals under EU review | | **WHO VigiAccess** | Global | Public view of the WHO global safety database | ### Clinical guidelines and standard of care How patients are treated today. | Source | Region | What it covers | | --- | --- | --- | | **HHS HIV guidelines** | US | Treatment guidelines: HIV | | **ACIP (CDC)** | US | Treatment guidelines: vaccines | | **Clinical practice guidelines from major medical societies** | Global | Treatment guidelines across therapy areas | ### Epidemiology and public health How many patients there are. | Source | Region | What it covers | | --- | --- | --- | | **SEER** | US | US cancer incidence, prevalence and survival | | **CDC WONDER** | US | US mortality and disease data | | **NHANES (NCHS)** | US | US national health and nutrition survey: obesity, diabetes, hypertension, and more | | **NHIS (NCHS)** | US | US national health interview survey | | **Eurostat health database** | EU | EU health and causes-of-death data | | **European Cancer Information System (ECIS)** | EU | EU cancer estimates by country | | **China CDC** | China | Chinese disease surveillance and reports | | **China National Cancer Center** | China | Annual Chinese cancer statistics | | **WHO Global Health Observatory** | Global | Global health indicators | | **OECD Health Statistics** | Global | Health and disease data for OECD countries | | **Orphanet prevalence data** | Global | Rare disease prevalence and incidence | | **UN World Population Prospects** | Global | Population by age and country | ### Pricing, reimbursement, and HTA What a drug costs, and whether it gets paid for. | Source | Region | What it covers | | --- | --- | --- | | **CMS ASP pricing files** | US | Medicare Part B average sales prices | | **NADAC (Medicaid)** | US | National average pharmacy acquisition costs by drug | | **Medicaid State Drug Utilization Data** | US | Medicaid prescriptions by drug and state | | **Medicare Part D Spending Dashboard** | US | Medicare Part D spend and claims by drug | | **Medicare Part B Spending Dashboard** | US | Medicare Part B spend and claims by drug | | **Medicare Drug Price Negotiation** | US | Drugs selected under the Inflation Reduction Act and negotiated prices | | **NICE technology appraisals** | UK | England reimbursement decisions | | **G-BA (AMNOG)** | EU | German benefit decisions that set pricing | | **IQWiG (AMNOG)** | EU | German benefit assessments behind G-BA decisions | | **HAS Transparency Committee** | EU | French benefit ratings | | **EU Joint Clinical Assessments** | EU | EU-wide HTA under the 2025 regulation | | **China National Reimbursement Drug List (NHSA)** | China | NRDL listings and negotiated prices | | **Japan NHI price listings** | Japan | Japanese reimbursement prices | ### Patents and IP Who owns what, and when exclusivity ends. | Source | Region | What it covers | | --- | --- | --- | | **US Patent and Trademark Office** | US | US patents and applications | | **European Patent Office** | EU | European patents and applications | | **Japan Patent Office** | Japan | Japanese patents and applications | | **China National Intellectual Property Administration** | China | Chinese patents and applications | | **WIPO PCT applications** | Global | International applications, often the first filing on a new program | | **Orange Book patents and exclusivities** | US | Listed patents and exclusivity expiry per drug | | **Paragraph IV certifications list** | US | Drugs facing generic patent challenges | | **Purple Book patent lists** | US | Biologic patents disclosed in biosimilar litigation | ### Company filings and disclosures What companies are required to disclose. | Source | Region | What it covers | | --- | --- | --- | | **SEC EDGAR** | US | 10-K, 10-Q, 8-K, S-1, F-1, 20-F, 6-K and proxy filings | | **SEC exhibits (license and collaboration agreements)** | US | Filed deal agreements with economics and terms | | **SEC ownership filings (13D, 13G, 13F)** | US | Activist stakes and institutional holdings | | **SEC Form D** | US | Private placements and venture rounds | | **HKEXnews** | China | Hong Kong listings, including 18A biotech prospectuses and announcements | | **CNInfo** | China | Shanghai and Shenzhen disclosures, including STAR Market biotechs | | **EDINET** | Japan | Japanese securities filings | | **TDnet** | Japan | Japanese timely disclosures | | **DART** | Korea | Korean securities filings | ### Company websites and investor materials What companies choose to say. | Source | Region | What it covers | | --- | --- | --- | | **Company pipeline pages** | Global | Programs, targets, modalities and phases as companies list them, public and private | | **Company press releases** | Global | Releases posted on company newsrooms | | **Corporate and investor presentations** | Global | Pipeline decks and corporate overviews | | **Earnings call transcripts** | Global | Management commentary and analyst Q&A | | **Investor day and R&D day materials** | Global | Slides and transcripts from strategy events | | **Healthcare investor conference presentations** | Global | Fireside chats and presentations at bank-hosted investor conferences and JPM week | | **Corporate brochures and white papers** | Global | Company-issued technical and platform documents | | **Posters and publications hosted on company sites** | Global | Scientific posters companies post after meetings | | **Company-hosted KOL events** | Global | Webcasts with key opinion leaders | | **Archived versions of company pages** | Global | Historical versions of pipeline pages, showing what was dropped or changed | | **Public market data** | Global | Share prices and market caps for listed companies, from public web sources | ### News and newswires What happened this week. | Source | Region | What it covers | | --- | --- | --- | | **Business Wire** | Global | Company press releases | | **PR Newswire** | Global | Company press releases | | **GlobeNewswire** | Global | Company press releases | | **Biotech and pharma trade press** | Global | Daily industry news on approvals, readouts, deals, and financings | ### Research funding and grants What academic science is funded before it becomes a company. | Source | Region | What it covers | | --- | --- | --- | | **NIH RePORTER** | US | NIH-funded projects, investigators and linked papers | | **ARPA-H** | US | US advanced health research programs and awards | | **SBIR / STTR awards** | US | US small business innovation grants to biotechs | | **EU CORDIS** | EU | Horizon Europe and ERC projects | ### Biology and disease reference data Public reference sets used alongside Sleuth's own ontology. | Source | Region | What it covers | | --- | --- | --- | | **Disease Ontology** | Global | Disease classification | | **MONDO** | Global | Unified disease ontology, strong on rare disease | | **MeSH** | Global | NLM's controlled vocabulary used across PubMed | | **ICD-10-CM** | US | US diagnosis codes | | **ICD-11** | Global | WHO international diagnosis codes | | **NCI Thesaurus** | US | Cancer terminology and drug names | | **Orphanet nomenclature** | Global | Rare disease names, genes and classifications | | **Open Targets** | Global | Target–disease evidence | | **UniProt** | Global | Protein sequences and function | | **HGNC** | Global | Official gene names | | **NCBI Gene** | Global | Gene records | | **Ensembl** | Global | Genome annotation | | **Human Protein Atlas** | Global | Tissue and cell expression | | **ChEMBL** | Global | Bioactive molecules and assay data | | **IUPHAR/BPS Guide to Pharmacology** | Global | Targets and ligands | | **PubChem** | Global | Chemical structures and bioactivity | | **FDA Global Substance Registration System (UNII)** | US | Unique substance identifiers | | **RxNorm** | US | Normalized US drug names | ### Social and alternative signals Early signals outside formal disclosure. | Source | Region | What it covers | | --- | --- | --- | | **WeChat official accounts** | China | Pipeline, financing and trial updates from Chinese biotechs, translated | | **Company job postings** | Global | Hiring by function and site, an early read on program and commercial build-out | | **People and executive moves** | Global | Leadership changes and team build-outs | | **Social posts from companies, KOLs, and analysts** | Global | Real-time commentary around data releases and meetings | | **Biotech podcasts and webinars** | Global | Executive and investor interviews, transcribed | ## Medical meetings Abstracts, posters, oral presentation slides, and late-breakers from about 365 medical meetings across 20 therapy areas, with archives back to 2015, including China-focused meetings. ## Sell-side research Research from 35+ banks and brokers. [Sell-side research](sell-side.md) ## Don't see a source you need? Sleuth adds sources when customers need them. [Tell us which one](get-help.md) ## Related - [What's in the knowledge base](knowledge-base.md): The records Sleuth works from, how many there are, and how they connect. - [China coverage](china.md): How Sleuth covers Chinese programs, and how they show up next to everyone else's. - [Sell-side research](sell-side.md): Research from 35+ banks and brokers, included in every plan, searchable next to everything else. *Updated November 9, 2026.* Source: md/sources.md --- # How fresh the data is How quickly new trials, filings, and deals reach Sleuth, and how to keep your own work current. [Open a project](https://app.sleuth.studio) ## The knowledge base Sleuth updates over 95% of its sources in real time. Every new document goes through the same matching and checks as everything else before it reaches the knowledge base, so a new record arrives already linked to the drugs and companies it's about. ## Your datasets A dataset is a snapshot of what Sleuth found when it built it. To keep one current, turn on monitoring: Sleuth updates the dataset as new trials, filings, and deals come in, shows you what changed, and keeps every earlier version. [Monitoring](monitoring.md) ## Quiet changes Sleuth keeps earlier versions of company pipeline pages, so a program dropped from a pipeline without a press release shows up as a change. Where there's a public reason a program stopped, Sleuth records it. ## Every value has a date A cell's source shows when the document was published, and recency counts toward the confidence score where it matters, like a trial's status. [Confidence scores](confidence.md) ## Related - [Keep a project current with monitoring](monitoring.md): Turn on monitoring and Sleuth updates your datasets as new trials, filings, and deals come in. - [Confidence scores and labels](confidence.md): How Sleuth scores every value, and what sourced, inferred, and unknown mean. - [Every source](sources.md): The registries, regulators, journals, filings, and feeds Sleuth reads, grouped by what they answer. *Updated November 9, 2026.* Source: md/data-freshness.md --- # Confidence scores and labels How Sleuth scores every value, and what sourced, inferred, and unknown mean. [Open a dataset](https://app.sleuth.studio) Every value in a Sleuth dataset links to its source and has a confidence score from 0 to 100. The score tells you how much to trust a value before you use it. ## How the score is worked out A fixed set of rules sets the score, so the same evidence always gets the same score and you can see why. Sleuth's model reads the sources and labels what each one says; the rules then score it on three things: - **Source quality.** A label, a registry, or a filing counts for more than a news story. - **Agreement.** Independent sources that agree raise the score. - **Recency,** where it matters, like a trial's status or a drug's phase. ## The labels | Score | Label | | --- | --- | | 85 to 100 | Confirmed | | 70 to 84 | High | | 50 to 69 | Moderate | | 25 to 49 | Low | | 0 to 24 | Unverified | ## Sourced, inferred, or unknown - **Sourced:** a document Sleuth links to states the value. - **Inferred:** Sleuth worked the value out from its sources and shows how. - **Unknown:** Sleuth couldn't find a public source, so it leaves the cell open instead of filling it in. Each dataset shows how many of its values fall in each group, and each column has its own score. ## When sources disagree Sleuth keeps both sources, goes with the one it scores higher, caps the cell's score, and flags it so you see the conflict. In Review mode you see both side by side and can switch to the other one. Your own corrections get a score of 100. ## How Sleuth avoids made-up values Sleuth checks every value before you see it. Its engineering matches and normalizes every record, AI checks run at each step, disagreements between sources are flagged instead of smoothed over, and Sleuth's data team tests the results. Because every value is labeled and linked, you can check any number in one click, and Review mode shows the few cells Sleuth couldn't settle, and why. [Review mode](review-mode.md) ## Related - [Check a dataset with Review mode](review-mode.md): See only the cells worth a second look, check their sources, and mark them reviewed. - [Datasets: rows, columns, and cells](datasets.md): What a Sleuth dataset is, which cells use credits, and how to change one. - [How fresh the data is](data-freshness.md): How quickly new trials, filings, and deals reach Sleuth, and how to keep your own work current. *Updated November 9, 2026.* Source: md/confidence.md --- # Check a dataset with Review mode See only the cells worth a second look, check their sources, and mark them reviewed. [Open a dataset](https://app.sleuth.studio) You don't have to check every cell. Values with strong sources that agree score high, and Review mode puts the rest in front of you, with the reason. ## Review a dataset 1. **Open the dataset** and switch from the dataset view to **Review mode**. 2. **Filter** with the flags: **Low confidence**, **Inferred**, and **Sources disagree**. Start with low confidence. 3. **Open a cell** to see its score, Sleuth's explanation, and its sources. Sources open inside Sleuth: papers, press releases, earnings transcripts, and broker research. 4. **Decide.** Mark the cell reviewed, flag it, leave a comment, or edit it by double-clicking. To have Sleuth try again, re-run the cell with guidance; each re-run uses one credit. ## When sources disagree For a disputed cell, Sleuth explains the conflict and suggests a value. Apply the suggestion or keep the current value. A common case: China's trial registry shows a Phase 3 start while ClinicalTrials.gov still lists Phase 2, and Sleuth used the newer record, which scored higher. ## Things to know - Your edits and review marks are saved with the dataset, and edited values get a confidence of 100. - You can undo a review mark. - You can't edit a past version of a dataset, or a dataset someone else owns. ## Related - [Confidence scores and labels](confidence.md): How Sleuth scores every value, and what sourced, inferred, and unknown mean. - [Datasets: rows, columns, and cells](datasets.md): What a Sleuth dataset is, which cells use credits, and how to change one. - [Run your first analysis](first-analysis.md): From sign-in to a finished analysis you can check and export, in five steps. *Updated November 9, 2026.* Source: md/review-mode.md --- # China coverage How Sleuth covers Chinese programs, and how they show up next to everyone else's. [See China data](https://www.sleuthinsights.com/china) Sleuth covers over 22,000 Chinese programs, as of November 2026, from China's own registries, regulators, exchanges, journals, company posts, and broker research. ## The Chinese sources - **Trial registries:** ChiCTR, the CDE drug trial registration platform, and ITMCTR. - **Regulators:** NMPA approvals and CDE application, clearance, and priority review lists. - **Exchanges:** HKEXnews, including 18A biotech prospectuses, and CNInfo. - **Literature:** Chinese-language journals, theses, and conference papers. - **Patents:** China National Intellectual Property Administration filings. - **Company posts:** WeChat official accounts of Chinese biotechs and industry media, where pipeline news often appears first. - **Pricing and statistics:** the National Reimbursement Drug List, China CDC, and the National Cancer Center. ## How it shows up Sleuth translates each source and runs it through the same matching and checks as every other source. A Chinese company's program appears in the same table as a US one, under the same drug record, and companies carry both their Chinese and English names. Search in English; you don't need to know the Chinese name. ## Related - [Every source](sources.md): The registries, regulators, journals, filings, and feeds Sleuth reads, grouped by what they answer. - [What's in the knowledge base](knowledge-base.md): The records Sleuth works from, how many there are, and how they connect. - [Explore the knowledge base](explore.md): Filter the records Sleuth works from and build your own dataset, with no credits. *Updated November 9, 2026.* Source: md/china.md --- # How credits work What a credit buys, what is free, and how to tell what a job will use before it runs. [See your balance](https://app.sleuth.studio) One credit is one analytical cell: a judgment Sleuth makes about one item in your research, like whether an asset fits your strategy, or how crowded an indication is for your modality. A credit costs about 6 to 8 cents, depending on your plan. [Plans and pricing](plans.md) ## Fact cells and analytical cells A **fact cell** repeats what a source states, like a drug's phase or sponsor, and it's free. An **analytical cell** is a judgment Sleuth makes for your question, like whether that drug fits your criteria, and it uses one credit. A dataset's plan shows which of its columns are analytical. ## What uses credits - **Analytical cells** in datasets Sleuth builds for your questions and workflows. - **Custom columns**: each cell Sleuth fills uses one credit. [Add your own columns](custom-columns.md) - **Re-running cells**, and adding analytical columns or rows to a dataset you already have. - **Checking the sources for a table you uploaded**: one credit per cell checked. [Bring your own documents and data](own-data.md) - **Monitoring**, but only when it finds something new and Sleuth analyzes it. [Monitoring](monitoring.md) ## What's free - Browsing the knowledge base and building datasets from it in Explore - Fact cells from the knowledge base, like phase, sponsor, and target - Searching and reading documents, broker research included - Asking about results you already have. A question that needs new rows or analytical columns runs a new plan, and those cells use credits. - Charts and exports - Runs that fail without results - Monitoring checks that find nothing new ## Worked examples | Job | Credits | | --- | --- | | Score 150 assets on 3 criteria | 450 | | Pull phase and sponsor for 100 drugs | 0 | | Add a column judging fit for each of 80 programs | 80 | | A landscape of 40 programs with 6 analytical columns and 8 fact columns | 240 | | Ask follow-up questions and make three charts from that landscape | 0 | ## Know before it runs Sleuth shows you its plan before it builds anything, so you see the size of the job first: how many rows and which columns. Rows times analytical columns is the credits it will use. The plan screen also shows your balance, and you can narrow the scope before you start. ## Estimate your team's month Count the jobs your team runs in a typical month and do the same math. If five people each build two landscapes of 150 rows with 3 analytical columns, that's 5 × 2 × 450 = 4,500 credits, which fits Team 6,000. On a trial, your balance shows what your own jobs use after a few days, which is the best guide. [Plans and pricing](plans.md) ## The same price everywhere A cell costs the same in Sleuth Studio as it does through Claude, ChatGPT, Copilot, or any other AI tool. Connecting an AI tool costs nothing extra. ## One balance per workspace On Team plans everyone draws from one pool. Admins see who used what on the [Usage page](usage-page.md). ## Related - [Plans and pricing](plans.md): Starter for one person, Team for any number of people, Enterprise for company-wide use, and what a credit costs. - [Check your balance and usage](balance-usage.md): Where to see your credits, how alerts work, and why the balance can lag. - [If you run out of credits](out-of-credits.md): What stops, what keeps working, and your options to carry on. *Updated November 9, 2026.* Source: md/how-credits-work.md --- # Plans and pricing Starter for one person, Team for any number of people, Enterprise for company-wide use, and what a credit costs. [See pricing](https://www.sleuthinsights.com/pricing) ## The plans | | Starter | Team | Enterprise | | --- | --- | --- | --- | | Price | $249 a month | From $499 a month | Talk to us | | Credits a month | 3,000 | 6,000, 10,000, 25,000, or 50,000, shared | Set in your agreement | | People | 1 | Any number | Any number | | Unused credits | Bank up to 2 months (6,000) | Monthly: bank up to 2 months. Annual: up to 2 extra months carry into the next year | Set in your agreement | | How you pay | Card, monthly | Card, monthly or annually. Invoice on annual plans from 10,000 credits a month | Invoice | | Analysis, the knowledge base, and document search | Yes | Yes | Yes | | Sell-side research from 35+ providers | Yes | Yes | Yes | | Sleuth in Claude, ChatGPT, and Copilot | Yes | Yes | Yes | | Workflows, your own documents, and monitoring | Yes | Yes | Yes | | Decks, reports, charts, and data exports | Yes | Yes | Yes | | Team workspace and admin analytics | No | Yes | Yes | | API, integrations, and Slack and Teams alerts | No | No | Yes | | SSO, security review, single tenant, and custom MSA | No | No | Yes | | Custom workflows and agents, built with your team | No | No | Yes | | Support | Email | Priority | White-glove onboarding | *Prices are in US dollars, before any sales tax.* ## Team prices | Credits a month | Monthly | Annual | | --- | --- | --- | | 6,000 | $499 a month | $416 a month, billed annually | | 10,000 | $799 a month | $665 a month, billed annually | | 25,000 | $1,899 a month | $1,583 a month, billed annually | | 50,000 | $3,599 a month | $2,999 a month, billed annually | Annual Team plans cost 10 times the monthly price, so you get two months free, and the year's credits arrive up front. ## What a credit costs Plans work out to about 6 to 8 cents a credit: 8.3 cents on Starter and Team 6,000, down to 7.2 cents on Team 50,000, and 6 to 7 cents on annual plans. Bigger plans and annual billing cost less per credit. ## Starter or Team? Starter is for one person. Team is for people who work together: everyone draws on one credit pool, and the admin sees who used what. Each person's projects are their own, and you share one by sending a copy. There's no per-seat price on Team, so adding a colleague costs nothing; only credits do. Each person signs in with their own account. ## Which Team size? Multiply the jobs your team runs in a month by the cells in each. Five people each building two landscapes of 150 rows with 3 analytical columns use about 4,500 credits a month, which fits Team 6,000. You can move up a size at any time, and the new credits arrive at once. [How credits work](how-credits-work.md) ## Bonuses Your first plan comes with an extra month of credits. Move up a size later, and the difference between plans comes as a bonus too. [Rollover and bonuses](rollover-bonuses.md) ## Free ways in - **The free trial:** 14 days, 2,500 credits, everything in Starter, no card. [Your free trial](free-trial.md) - **Sleuth for Startups:** a free year for companies that have raised under $10 million. [Sleuth for Startups](startups.md) ## Plans cover your own team's work To build Sleuth data into a product you sell, you'll need Enterprise. [Enterprise](enterprise.md) ## Related - [How credits work](how-credits-work.md): What a credit buys, what is free, and how to tell what a job will use before it runs. - [Buy a plan](buy-plan.md): Choose a plan, pay by card, and what you get when the payment goes through. - [Enterprise](enterprise.md): What Enterprise adds for companies rolling Sleuth out across teams and systems. *Updated November 9, 2026.* Source: md/plans.md --- # Check your balance and usage Where to see your credits, how alerts work, and why the balance can lag. [Open Studio](https://app.sleuth.studio) ## Where to see your credits - **The sidebar card** shows the credits you have left and when the next credits arrive. - **Project and dataset headers** show your remaining credits. Hover for **About your credits**. - **The Usage page** in Settings shows what used them, for admins. [Usage page](usage-page.md) On Team plans, members see the balance only if an admin turns on **Show credit availability to all users** on the Usage page. ## Credit alerts Sleuth emails your workspace admin when the workspace has used 75% and 90% of its credits, and again when they run out. ## Why the balance lags Usage reaches your balance in batches, so the number you see can trail what you've used by a few hours. A run that's already going always finishes, even if it takes the balance a little past zero. ## When new credits arrive Monthly plans add the month's credits on your billing date. Annual plans add the year's credits when you buy and on each renewal. Unused credits roll over up to a limit. [Rollover and bonuses](rollover-bonuses.md) ## Related - [The Usage page](usage-page.md): See what used your credits, by event, project, and person, and export it. - [If you run out of credits](out-of-credits.md): What stops, what keeps working, and your options to carry on. - [Rollover and bonuses](rollover-bonuses.md): How unused credits carry over, which credits never expire, and which credits Sleuth uses first. *Updated November 9, 2026.* Source: md/balance-usage.md --- # If you run out of credits What stops, what keeps working, and your options to carry on. [See plans](https://app.sleuth.studio/settings/billing) ## What happens - **Nothing is lost.** Your projects, datasets, and chats stay where they are. - **Runs already going finish.** Sleuth checks your balance before a run starts, so nothing stops halfway. - **New credit actions pause.** Workflows, questions that need new analysis, custom columns, and monitoring wait until you have credits. - **Downloads and exports pause** until you have credits again. - **Everything else works.** You can open projects, read results, browse Explore, and search documents. ## What you see | You are | Sleuth shows | Your options | | --- | --- | --- | | On the trial | **You've used your trial credits** | See plans (Starter $249 a month, Team from $499), or talk to us | | The admin of a paid plan | **You're out of credits for this month** | Upgrade your plan now, or wait until your next credits arrive | | A member of a Team workspace | **Your team is out of credits** | Ask your admin to upgrade | Anything you typed in the chat box stays there, so you can send it once credits are back. ## Get more credits - **Upgrade.** An upgrade takes effect right away and adds the new plan's credits at once. Your unused credits stay. [Change your plan](change-plan.md) - **Wait for your next credits.** Monthly plans add credits on your billing date. - **Talk to us** if you need a one-off top-up for a big project. [Get help](get-help.md) ## Extra usage Extra usage is coming in December. A monthly spending limit will let you keep going past your plan's credits at a fixed price per credit, up to the limit you set. Every workspace's limit will start at $0, so nothing extra is charged unless an admin raises it. ## Related - [Check your balance and usage](balance-usage.md): Where to see your credits, how alerts work, and why the balance can lag. - [Change, downgrade, or cancel your plan](change-plan.md): Upgrades happen now; downgrades and cancellations wait for the end of your billing period. Also how refunds work. - [How credits work](how-credits-work.md): What a credit buys, what is free, and how to tell what a job will use before it runs. *Updated November 9, 2026.* Source: md/out-of-credits.md --- # Buy a plan Choose a plan, pay by card, and what you get when the payment goes through. [Pick a plan](https://app.sleuth.studio/settings/billing) ## Choose and pay 1. **Open the plans.** Press **Upgrade** on the sidebar card, or open Settings, then Billing. You'll see **Pick a plan** with Starter, Team, and Enterprise, and a Monthly and Annual switch. Annual shows Team only. 2. **Choose a plan.** For Team, pick the size: 6,000, 10,000, 25,000, or 50,000 credits a month. 3. **Enter your billing address.** Sleuth needs it to work out sales tax. Press **Continue to payment**. 4. **Pay on the secure payment page** by card, in US dollars. If your bank asks you to approve the payment, finish the check there. Card details go to the payment processor and never reach Sleuth. 5. **Wait a few seconds.** You'll see **Setting up your plan**, then a confirmation. ## Buying from the pricing page If you pick a plan on [sleuthinsights.com/pricing](https://www.sleuthinsights.com/pricing), you go straight to checkout, without a trial. Leave checkout before paying, and Studio shows **Finish setting up your plan**, with **Continue to checkout**, **See all plans**, and **Start a free trial instead**. ## What you get - **Your plan's credits**, for the month or, on annual plans, for the year. - **Your unused trial credits**, which carry over and never expire. - **Your welcome bonus:** an extra month of credits. [Rollover and bonuses](rollover-bonuses.md) - **Admin rights.** The person who pays becomes the workspace admin. On Team, the confirmation shows **Invite teammates**, so you can bring colleagues into the same credit pool. [Invite your team](invite-team.md) ## If something goes wrong - **You leave the payment page without paying.** Nothing changes and nothing is charged. Your workspace stays as it was. - **Your card is declined.** Sleuth shows **Your card was declined. Change the card and try again.** A declined attempt isn't charged. Try another card, or ask your bank to allow the payment. - **You paid but the plan hasn't started.** Sleuth shows **Your payment went through, but your plan didn't start.** Wait a minute and reload. If it still hasn't, [contact support](get-help.md); your payment is safe. ## Paying by invoice Annual Team plans of 10,000 credits a month or more, and Enterprise, can pay by invoice and bank transfer, within 30 days. For purchase orders or vendor forms like a W-9, email sales@sleuthinsights.com, or [talk to us](https://www.sleuthinsights.com/talk-to-us). ## Related - [Plans and pricing](plans.md): Starter for one person, Team for any number of people, Enterprise for company-wide use, and what a credit costs. - [Change, downgrade, or cancel your plan](change-plan.md): Upgrades happen now; downgrades and cancellations wait for the end of your billing period. Also how refunds work. - [Invoices, cards, and payment problems](payment-problems.md): Change your card, find invoices, and what happens if a renewal payment fails. *Updated November 9, 2026.* Source: md/buy-plan.md --- # Change, downgrade, or cancel your plan Upgrades happen now; downgrades and cancellations wait for the end of your billing period. Also how refunds work. [Open Billing](https://app.sleuth.studio/settings/billing) Admins change plans in Settings, then Billing. Every change shows a preview before you confirm. ## Upgrade An upgrade takes effect right away: - You pay the new plan's price now, less a credit for the unused part of your current plan. The preview shows the exact amount. - Your billing date moves to today. - The new plan's credits arrive at once, and your unused credits stay. For example, moving from Starter to Team 6,000 on the day you bought Starter costs about $249 before tax, because Sleuth credits almost all of the Starter month. Switching from monthly to annual is an upgrade too: open the plans from Billing and turn on Annual. ## Downgrade A downgrade takes effect at the end of your billing period, or at the end of the year on an annual plan. Nothing is charged now, and you keep your current plan and its credits until then. A banner shows the pending change; press **Keep** on it to cancel the downgrade. Starter is for one person, so to move from Team to Starter, remove the other members first. ## Cancel Press **Cancel plan** and confirm. There's no cancellation fee or minimum term on monthly plans. Your plan runs until the end of the period you paid for, then the workspace becomes view only: your projects stay and you can read them, but chat and new runs pause. Keep the plan from the banner if you change your mind before then. To start again later, choose a plan. Credits that never expire, like trial and bonus credits, are still there when you come back. *Billing dates show in UTC, so in the Americas a renewal can look a day early.* ## Refunds Sleuth's terms of service allow a refund of a renewal if you cancel within 15 days of it. If you were charged by mistake, for example by an upgrade you didn't mean to make, email support@sleuthinsights.com right away and Sleuth will sort it out. ## Related - [Buy a plan](buy-plan.md): Choose a plan, pay by card, and what you get when the payment goes through. - [Rollover and bonuses](rollover-bonuses.md): How unused credits carry over, which credits never expire, and which credits Sleuth uses first. - [Invoices, cards, and payment problems](payment-problems.md): Change your card, find invoices, and what happens if a renewal payment fails. *Updated November 9, 2026.* Source: md/change-plan.md --- # Rollover and bonuses How unused credits carry over, which credits never expire, and which credits Sleuth uses first. [See your balance](https://app.sleuth.studio) ## Rollover - **Monthly plans** bank up to two months of credits. On Starter that's 6,000: a month you don't use carries into the next, and anything older than two months expires. - **Annual plans** carry up to two extra months of credits into the next year, on top of the year's credits. - **Bonus and trial credits never expire** and don't count toward either limit. ## Bonuses - **Welcome bonus.** Your first plan comes with an extra month of credits. - **Upgrade bonus.** Move up a size later, and the difference in monthly credits between the two plans comes as a bonus once you've paid for a full month on the bigger plan. Each bonus is given once per plan, Starter or Team. ## Which credits Sleuth uses first Sleuth spends the credits that would expire soonest first: last month's plan credits, then this month's, then trial and bonus credits, which never expire. That way as few credits as possible go unused. ## Related - [Plans and pricing](plans.md): Starter for one person, Team for any number of people, Enterprise for company-wide use, and what a credit costs. - [Check your balance and usage](balance-usage.md): Where to see your credits, how alerts work, and why the balance can lag. - [Change, downgrade, or cancel your plan](change-plan.md): Upgrades happen now; downgrades and cancellations wait for the end of your billing period. Also how refunds work. *Updated November 9, 2026.* Source: md/rollover-bonuses.md --- # Invoices, cards, and payment problems Change your card, find invoices, and what happens if a renewal payment fails. [Open Billing](https://app.sleuth.studio/settings/billing) ## Change your card Open Settings, then Billing, and press **Change card**. Add the new card on the secure page and return to Sleuth. Billing shows the new card's brand and last four digits, and future payments use it. ## Invoices and receipts Press **View invoices** in Billing to see every invoice and payment. Invoices and receipts are emailed to the person who pays, from Orb, Sleuth's billing provider, so search your inbox for that name. Next payment, shown in Billing, is before tax. ## Sales tax Prices are in US dollars. Sleuth adds sales tax based on your billing address. To change the address, contact support. ## If a renewal payment fails 1. **You get an email from Orb,** and Sleuth retries the card 1, 4, and 9 days later. Until the last retry, the admin sees **We couldn't charge your card for [plan]. We'll try again on [date].** with **Update payment**. Nothing is blocked, and members see nothing. 2. **Add a new card** in Billing at any point, and Sleuth retries right away. 3. **If the last retry fails,** everyone sees a red banner and new work waits for the payment. Members see **Ask your admin to update payment.** A new card unpauses the workspace within seconds, and your projects stay throughout. ## Paying by invoice Annual Team plans of 10,000 credits a month or more, and Enterprise, can pay by invoice and bank transfer, within 30 days. For purchase orders or vendor forms like a W-9, email sales@sleuthinsights.com, or [talk to us](https://www.sleuthinsights.com/talk-to-us). ## Related - [Buy a plan](buy-plan.md): Choose a plan, pay by card, and what you get when the payment goes through. - [Change, downgrade, or cancel your plan](change-plan.md): Upgrades happen now; downgrades and cancellations wait for the end of your billing period. Also how refunds work. - [Notifications](notifications.md): The emails Sleuth sends and how to change them. *Updated November 9, 2026.* Source: md/payment-problems.md --- # Sleuth for Startups A free year of Sleuth for your whole team, at no cost and with no card, if your company has raised under $10 million. [Apply](https://www.sleuthinsights.com/startups) ## Who qualifies A life-science company that has raised under $10 million in total, including companies that haven't raised yet. Each company applies for itself, so investors can send the link to their portfolio companies. Investors and incubators that back early companies can apply too, and their own team gets Sleuth free for a year. ## What you get - **12 months free:** $0, with no card needed. - **A shared workspace for your whole team,** with 3,000 credits a month that everyone draws on. - **Starter's features:** workflows, the knowledge base, document search, sell-side research, and Sleuth in Claude, ChatGPT, and Copilot. ## How to apply 1. **Apply** on [sleuthinsights.com/startups](https://www.sleuthinsights.com/startups). Tell us about your company and the first question you'd bring to Sleuth. 2. **Sleuth reviews every application** and emails you when your team's account is ready. You can start the 14-day free trial in the meantime. 3. **Bring your team.** Invite colleagues, and everyone shares the same credits. ## After 12 months Choose Starter ($249 a month for one person) or Team (from $499 a month for any number of people), or stop. Your projects stay in your account either way. [Plans and pricing](plans.md) ## Raised more than $10 million? Start the 14-day free trial instead. [Your free trial](free-trial.md) ## Related - [Your free trial](free-trial.md): 14 days and 2,500 credits with everything in Starter, no card needed. - [Plans and pricing](plans.md): Starter for one person, Team for any number of people, Enterprise for company-wide use, and what a credit costs. - [Invite your team](invite-team.md): Bring colleagues into a Team workspace so everyone draws on one pool of credits, fix an invite that didn't arrive, and remove someone who left. *Updated November 9, 2026.* Source: md/startups.md --- # Invite your team Bring colleagues into a Team workspace so everyone draws on one pool of credits, fix an invite that didn't arrive, and remove someone who left. [Invite teammates](https://app.sleuth.studio) Inviting colleagues needs a Team or Enterprise plan. The trial and Starter are for one person. On those, the invite step shows **Inviting teammates is on Team**, with a link to the Team plans. ## Invite by email 1. Press **Invite teammates**, on the confirmation after you buy Team or in Settings. 2. Enter their email addresses and send. Each person gets an email titled **[your name] invited you to Sleuth**. 3. They join from the email and finish onboarding. An invited personal email address, like Gmail, still waits for Sleuth's approval. ## Colleagues who sign up on their own When someone signs up with your company's email domain, onboarding offers **Join your team's workspace** with your workspace's name, so they land in the same place. ## What members can do Members run workflows, ask questions, and build datasets, all drawing on the shared credits. Each person's projects are their own. To work on something together, send a copy your colleague can keep working in, or transfer it to them. [Share your work](share-work.md) Members don't see Billing or upgrade links. [Admins and members](roles.md) ## If an invite didn't work - **Check the plan.** Invites work on Team and Enterprise only. - **Check the date.** Invites expire after 30 days. Send a new one. - **Check spam** for the email titled **[your name] invited you to Sleuth**. - **Personal email addresses** wait for Sleuth's approval before they can join. - Still stuck? [Contact support](get-help.md) with the address you invited. ## Remove someone To remove a member, [contact support](get-help.md) with their email address, and Sleuth removes them from the workspace. ## Related - [Admins and members](roles.md): What a workspace admin controls and what members see. - [Projects, initiatives, and sharing](share-work.md): How projects are organized, who can see them, and how to get work to people outside Sleuth. - [Plans and pricing](plans.md): Starter for one person, Team for any number of people, Enterprise for company-wide use, and what a credit costs. *Updated November 9, 2026.* Source: md/invite-team.md --- # Admins and members What a workspace admin controls and what members see. [Open Studio](https://app.sleuth.studio) ## Who's the admin The person who buys the plan becomes the workspace admin. On a trial, the person who started it manages billing. To add an admin or move admin rights to someone else, [contact support](get-help.md). ## What admins can do | Area | Admins | Members | | --- | --- | --- | | Run workflows, ask questions, build datasets | Yes | Yes | | Send a copy of a project, or transfer it | Yes | Yes | | See the plan, card, and invoices; change or cancel the plan | Yes | No | | See usage by person and project | Yes | No | | See the credit balance | Yes | If an admin turns it on | | Invite members | Yes | No | | Remove members | Through support | No | ## Spending per person Everyone on a Team plan draws on the same pool of credits, and there's no cap per person. The [Usage page](usage-page.md) shows who used what, so admins can see where the credits go. ## Show credits to everyone On the Usage page, turn on **Show credit availability to all users**. Members then see the balance in the sidebar and in project headers within a few minutes. ## When the team runs out Admins see **You're out of credits for this month**, with options to upgrade or wait. Members see **Your team is out of credits**. [If you run out of credits](out-of-credits.md) ## Related - [Invite your team](invite-team.md): Bring colleagues into a Team workspace so everyone draws on one pool of credits, fix an invite that didn't arrive, and remove someone who left. - [The Usage page](usage-page.md): See what used your credits, by event, project, and person, and export it. - [Change, downgrade, or cancel your plan](change-plan.md): Upgrades happen now; downgrades and cancellations wait for the end of your billing period. Also how refunds work. *Updated November 9, 2026.* Source: md/roles.md --- # The Usage page See what used your credits, by event, project, and person, and export it. [Open Usage](https://app.sleuth.studio) Admins open the Usage page from Settings. ## What it shows - **Your balance:** the percent of credits left, "X of Y credits left," and a bar marked at 75% and 90%. - **Usage by event:** projects created, datasets built, cells filled, and outputs made. - **Usage by project** and **by person**, so you can see where the credits went. ## Filter and export Pick a range: today, the last 7 days, the last 30 days, year to date, all time, or custom dates. Press export to download the report as an Excel file. ## Share the balance Turn on **Show credit availability to all users** to let members see the balance. It's off by default. ## Need more? The page links to request credits or talk to sales. ## Related - [Check your balance and usage](balance-usage.md): Where to see your credits, how alerts work, and why the balance can lag. - [Admins and members](roles.md): What a workspace admin controls and what members see. - [How credits work](how-credits-work.md): What a credit buys, what is free, and how to tell what a job will use before it runs. *Updated November 9, 2026.* Source: md/usage-page.md --- # Enterprise What Enterprise adds for companies rolling Sleuth out across teams and systems. [Talk to us](https://www.sleuthinsights.com/talk-to-us) Enterprise includes everything in Team, plus: ## Connect Sleuth to your systems - **API access** for applications and systems that call Sleuth directly. [API](api.md) - **Integrations** with Drive, SharePoint, Box, and your data warehouse, with your internal program names matched to Sleuth's records. - **Alerts in Slack and Microsoft Teams.** ## Security and procurement - **Single sign-on** through your identity provider, with your own sign-in and MFA rules. Sleuth sets it up with your IT team. - **Support for your security review,** including your questionnaire, and vendor paperwork like a W-9. - **A single-tenant option:** a dedicated cloud account, your own Studio and API backend, and private storage, on the same knowledge base. - **A custom MSA.** - **Your work stays separate:** your company's questions, documents, and results are never pooled with other customers' data, and Sleuth deletes them on request. ## Help rolling it out - **White-glove onboarding:** Sleuth's team sets up each group and trains it on its own work. - **Custom workflows and agents,** built with your team for your use cases. - **Early access to proprietary signal data:** which targets and indications are gaining attention across the industry. ## Pricing Enterprise has no list price. Sleuth prices it for the teams, systems, and security requirements involved, and Enterprise pays by invoice. [Talk to us](https://www.sleuthinsights.com/talk-to-us) ## Related - [Security overview](security.md): The short answers for your security team, and where to find the evidence. - [API](api.md): Sleuth’s REST API lets your own software run Sleuth research without the Studio screen. It comes with Enterprise and powers the MCP server. - [Plans and pricing](plans.md): Starter for one person, Team for any number of people, Enterprise for company-wide use, and what a credit costs. *Updated November 9, 2026.* Source: md/enterprise.md --- # Security overview The short answers for your security team, and where to find the evidence. [Visit the Trust Center](https://trust.sleuthinsights.com) | Question | Answer | | --- | --- | | Is Sleuth SOC 2 compliant? | Yes. Sleuth is audited under SOC 2 Type II, and the current report is from May 2026. Email security@sleuthinsights.com for a copy. | | Is our data encrypted? | Yes, in transit and at rest. | | Where is it stored? | In the United States. | | How is the network protected? | Sleuth runs in private networks (VPCs) in the cloud, protected by industry-standard configuration and practices. Network security, access control, vulnerability management, and incident response are among the controls audited under SOC 2 Type II. Sleuth's latest penetration test was in November 2026. | | Is our work private? | Yes. Your questions, uploads, and results stay in your workspace and are never shown to other customers. [How your data is used](data-use.md) | | Which AI models process our data? | Large language models from Anthropic, OpenAI, and Google, hosted by those companies. On Enterprise, you can choose models hosted on AWS Bedrock instead, so your data stays in Sleuth's private network and isn't sent to the model providers. | | Which other companies handle our data? | Email security@sleuthinsights.com for the current list of subprocessors. | | How do people sign in? | With an email and password, or a one-time code sent to their email. On Enterprise, people can also sign in through your identity provider with SSO. Sleuth can turn on multi-factor sign-in for your whole organization on request. | | Can we get our own environment? | On Enterprise, yes: the single-tenant option gives you a dedicated cloud account, your own Studio and API backend, and private storage. | | Can you fill in our security questionnaire? | Yes, on Enterprise, along with a custom MSA. On any plan, your security team can email security@sleuthinsights.com with questions. | | Can we upload patient data? | No. Sleuth's terms don't allow protected health information. | | Can we have our data deleted? | You can delete your own projects and documents from your workspace at any time. To have your personal data deleted, email privacy@sleuthinsights.com. On Enterprise, Sleuth deletes your company's questions, documents, and results on request. | ## The evidence The [Trust Center](https://trust.sleuthinsights.com) lists Sleuth's security controls and answers common questions. For the SOC 2 report, the subprocessor list, or anything else, email security@sleuthinsights.com. ## Card payments Card details go straight to Sleuth's payment processor and never reach Sleuth's systems. ## Related - [How your data is used](data-use.md): Who can see your questions, uploads, and results, and how to delete them. - [Enterprise](enterprise.md): What Enterprise adds for companies rolling Sleuth out across teams and systems. - [Your data in AI tools](ai-tools-data.md): What travels between your AI tool and Sleuth when you connect them. *Updated November 9, 2026.* Source: md/security.md --- # How your data is used Who can see your questions, uploads, and results, and how to delete them. [Read the privacy policy](https://www.sleuthinsights.com/privacy-policy) ## Your work is private On every plan, your questions, uploads, and results stay in your workspace and are never shown to other customers. Each project belongs to the person who made it. ## Uploads Files you upload are used for your work in your workspace. Sleuth reads them alongside its own sources when it answers your questions. They aren't added to the knowledge base other customers use. Don't upload patient health information: Sleuth's terms don't allow it. ## Deleting your data - Delete any project or uploaded document from your workspace, at any time. - To close your account and have your personal data deleted, email privacy@sleuthinsights.com. - On Enterprise, Sleuth deletes your company's questions, documents, and results on request. ## The details The [privacy policy](https://www.sleuthinsights.com/privacy-policy) and [terms of service](https://www.sleuthinsights.com/terms-of-service) have the full terms. ## Related - [Security overview](security.md): The short answers for your security team, and where to find the evidence. - [Your data in AI tools](ai-tools-data.md): What travels between your AI tool and Sleuth when you connect them. - [Use your own documents and data](own-data.md): Upload files for Sleuth to read, start from your own spreadsheet, or have Sleuth source the values you already have. *Updated November 9, 2026.* Source: md/data-use.md --- # Your data in AI tools What travels between your AI tool and Sleuth when you connect them. [Set up Sleuth in your AI tool](https://www.sleuthinsights.com/mcp-setup) ## Sign-in You sign in to Sleuth on Sleuth's own sign-in page and choose **Allow**. Your AI tool never sees your sign-in details; it gets a token that lets it act as you in Sleuth until you disconnect it. ## What the AI tool sends to Sleuth - Your name and email address, from sign-in. - The questions you ask Sleuth to work on, and your replies, like approving a plan. - Project names, search terms, and documents you choose to upload. - If you set up your account from the AI tool, the answers you confirm: your job title, the jobs you want done, how you work, and where you'll use Sleuth. If you request access there, also your company, country, and company website or LinkedIn profile. Each Sleuth tool takes only the fields its call needs. Sleuth's tools don't read your AI tool's memory, your chat history, or your files. ## You approve each tool Claude asks before it uses a Sleuth tool, unless you set Sleuth's tools to **Always allow**. Tools that rename, delete, cancel, or reject always ask. ## Sites to allow If your company filters web traffic, IT may need to allow these: | Address | What it's for | | --- | --- | | `mcp.sleuthapi.com` | The MCP server your AI tool connects to | | `auth.sleuth.studio` | Sign-in | | `app.sleuth.studio` | Sleuth Studio, where project links open | | `public-api.sleuth.studio` | Dataset downloads | | `storage.googleapis.com` | Research files Sleuth hosts | ## What Sleuth sends back Answers, datasets, charts, document passages, and links. Links point to Sleuth Studio, to dataset downloads, to Sleuth-hosted research files, and to publishers' pages for outside documents. ## Where it's kept Projects and documents you create through an AI tool are saved in your Sleuth account like any other Sleuth work, under the same privacy rules. [How your data is used](data-use.md) ## Disconnect Remove or disconnect Sleuth in your AI tool's connector settings. Your projects stay in your Sleuth account. ## Related - [Sleuth in Claude, ChatGPT, and Copilot](ai-tools-overview.md): Ask in the AI tool you already use, and Sleuth does the work behind the answer. - [How your data is used](data-use.md): Who can see your questions, uploads, and results, and how to delete them. - [Security overview](security.md): The short answers for your security team, and where to find the evidence. *Updated November 9, 2026.* Source: md/ai-tools-data.md --- # Sleuth in Claude, ChatGPT, and Copilot Ask in the AI tool you already use, and Sleuth does the work behind the answer. [Set it up](https://www.sleuthinsights.com/mcp-setup) Sleuth connects to Claude, ChatGPT, Microsoft Copilot, and any AI tool that supports MCP, the standard way AI tools connect to outside data. Once it's connected, your AI tool can hand Sleuth a question, and Sleuth builds the analysis from its knowledge base and documents, with a source on every value. ## Why connect Sleuth On its own, an AI tool searches the web and summarizes what it finds, and it can miss programs or get values wrong without saying where. With Sleuth connected, it works from Sleuth's knowledge base: drugs, trials, deals, and companies already matched and checked, with a source and a confidence score on every value, plus sell-side research and Chinese sources the open web doesn't reach. In Sleuth's benchmark, AI tools found 63% more with Sleuth connected. [See the benchmark](https://www.sleuthinsights.com/benchmarks) ## What you can do from your AI tool - **Run research projects.** Ask a question; Sleuth plans it, builds the datasets, and reports back. Every project is saved to your Studio account, where you or your team can open it and check its sources. - **Look things up.** Search companies, drugs, indications, trials, and deals, and find drugs by target, mechanism, modality, or therapeutic area. - **Search documents.** Search the research library, broker research included, and pull the passages that answer your question. - **Build datasets from the knowledge base** and get a download link. - **Work with what you already have.** List your projects, read datasets row by row or as a CSV, fetch charts, and upload documents. ## What happens in Studio Workflows, monitoring, sharing, and plan changes happen in Studio. Your AI tool can ask for the same analysis as a workflow in plain words. [Prompts to start with](ai-prompts.md) ## Before you start - **You need a Sleuth account.** Any plan works, including the free trial. If you don't have one, your AI tool can request access for you after you connect; you'll finish two quick steps in the browser, then reconnect. - **It costs nothing extra.** A cell costs the same credits in your AI tool as in Studio, and your AI tool only calls Sleuth when a question needs it. - **Plans may pause for approval.** Unless you turned on auto-approve in Studio, research runs wait for you to approve the plan in the chat. [Plans and approvals](plans-approvals.md) ## Set up your tool - [Claude](setup-claude.md): claude.ai, Claude Desktop, and Claude Code - [ChatGPT](setup-chatgpt.md): Add Sleuth as a connector - [Microsoft Copilot](setup-copilot.md): Through Copilot Studio - [Other AI tools](setup-other.md): Cursor, Gemini, and your own agents ## The server URL ``` https://mcp.sleuthapi.com/mcp ``` Every AI tool uses the same URL. Opening it in a browser shows "Method not allowed," which is expected. ## When to use Studio instead Use your AI tool for questions where you already are. Use Studio for projects you'll keep working on, to review values cell by cell, to run workflows, and to manage monitoring. Both work on the same projects. ## Related - [Set up Sleuth in Claude](setup-claude.md): Connect Sleuth on claude.ai, Claude Desktop, and Claude Code. - [Prompts to start with](ai-prompts.md): Fill in the brackets and paste a prompt into your AI tool. Sleuth shows you its plan before it builds anything. - [Your data in AI tools](ai-tools-data.md): What travels between your AI tool and Sleuth when you connect them. *Updated November 9, 2026.* Source: md/ai-tools-overview.md --- # Set up Sleuth in Claude Connect Sleuth on claude.ai, Claude Desktop, and Claude Code. [Open Claude](https://claude.ai) ## claude.ai and Claude Desktop Add Sleuth as a custom connector. It takes about a minute. 1. **Copy the server URL.** `https://mcp.sleuthapi.com/mcp` 2. **Add Sleuth to Claude.** In Claude, open **Customize**, then **Connectors**. Press **+ Add**, choose **Add custom connector**, name it Sleuth, and paste the URL. 3. **Approve access.** Claude opens a Sleuth sign-in page. Sign in with your Sleuth account and choose **Allow**. Claude never sees your sign-in details. 4. **Turn Sleuth on** in a chat from the tools menu, and ask your question. On a Team or Enterprise plan, an owner may need to add the connector for your organization first. Connectors are account-wide, so once Sleuth is connected it's available on claude.ai, Claude Desktop, and the Claude mobile apps. Claude asks before it uses each Sleuth tool. To skip the prompts for Sleuth's read-only tools, set them to **Always allow** in the connector's settings. Tools that rename, delete, cancel, or reject always ask. Sleuth's listing in Claude's connector directory is coming soon. Once it's live, you'll be able to find Sleuth there and press **Connect** instead. ## Claude Code 1. **Add the server** from your terminal: `claude mcp add --transport http sleuth https://mcp.sleuthapi.com/mcp` 2. **Sign in.** In Claude Code, run `/mcp`, choose sleuth, and sign in with your Sleuth account in the browser window that opens. To use Sleuth in every project on your machine, not only the current one, add it with `--scope user`: ``` claude mcp add --transport http --scope user sleuth https://mcp.sleuthapi.com/mcp ``` Or install the Sleuth plugin, which adds the same server plus a skill that helps Claude word research questions, follow long runs, and wait for your approval: ``` claude plugin marketplace add SleuthInsights/sleuth-plugins claude plugin install sleuth-insights@sleuth-plugins ``` Then run `/mcp`, choose sleuth, and sign in. If Sleuth stops answering later, run `/mcp` again and sign in. ## Check it worked Ask Claude: "List my Sleuth projects." A list, or a note that you have none yet, means you're connected. If Claude answers without using Sleuth, see [Fix connection problems](ai-troubleshooting.md). ## Try it Start with a prompt built for a common job. [Prompts to start with](ai-prompts.md) ## Related - [Prompts to start with](ai-prompts.md): Fill in the brackets and paste a prompt into your AI tool. Sleuth shows you its plan before it builds anything. - [Fix connection problems](ai-troubleshooting.md): The most common problems connecting Sleuth to an AI tool, and how to fix each one. - [Sleuth in Claude, ChatGPT, and Copilot](ai-tools-overview.md): Ask in the AI tool you already use, and Sleuth does the work behind the answer. *Updated November 9, 2026.* Source: md/setup-claude.md --- # Set up Sleuth in ChatGPT Add Sleuth to ChatGPT as a custom connector with the server URL. [Open ChatGPT](https://chatgpt.com) Add Sleuth as a custom connector. It takes about a minute. 1. **Copy the server URL.** `https://mcp.sleuthapi.com/mcp` 2. **Add Sleuth to ChatGPT.** In ChatGPT, open **Plugins** and press **+**. Choose **Add custom MCP server**, paste the URL, and choose OAuth. 3. **Approve access.** ChatGPT opens a Sleuth sign-in page. Sign in with your Sleuth account and choose **Allow**. 4. **Turn Sleuth on** in a chat and ask your question. ## On a company workspace On a Business or Enterprise workspace, an admin may need to allow custom connectors first. ## Once Sleuth is listed Sleuth's listing in ChatGPT's app directory is coming soon. Once it's live, you'll be able to search for Sleuth in ChatGPT and press **Connect** instead. ## Check it worked Ask: "Use Sleuth to list my projects." A list, or a note that you have none yet, means you're connected. ## Related - [Prompts to start with](ai-prompts.md): Fill in the brackets and paste a prompt into your AI tool. Sleuth shows you its plan before it builds anything. - [Fix connection problems](ai-troubleshooting.md): The most common problems connecting Sleuth to an AI tool, and how to fix each one. - [Set up Sleuth in Claude](setup-claude.md): Connect Sleuth on claude.ai, Claude Desktop, and Claude Code. *Updated November 9, 2026.* Source: md/setup-chatgpt.md --- # Set up Sleuth in Microsoft Copilot Add Sleuth to an agent in Copilot Studio, then use it in Microsoft 365 Copilot or Teams. [Open Copilot Studio](https://copilotstudio.microsoft.com) Sleuth works in Microsoft 365 Copilot through Copilot Studio. Someone with access to Copilot Studio adds Sleuth to an agent once, then publishes that agent where your team works. Your team then chats with that agent in Microsoft 365 Copilot or Teams, and the agent calls Sleuth when a question needs it. If your company's Copilot is locked down, IT decides whether custom tools like this are allowed. ## Who does what | Who | What they do | | --- | --- | | The agent builder, with Copilot Studio access | Adds Sleuth to an agent, tests the sign-in, and publishes the agent (the steps below) | | IT | Allows custom connectors or MCP tools in Copilot Studio, allows publishing to Microsoft 365 Copilot or Teams, and allows Sleuth's sign-in page and addresses | | Each person on your team | Signs in with their own Sleuth account the first time they use the agent | ## Add Sleuth to an agent 1. **Open your agent** in Copilot Studio and go to Tools. 2. **Add a Model Context Protocol tool.** Name it Sleuth and paste the server URL. 3. **Choose OAuth 2.0** with dynamic discovery as the authentication type. 4. **Sign in** with your Sleuth account to test the connection. 5. **Publish the agent** to Microsoft 365 Copilot or Teams. ``` https://mcp.sleuthapi.com/mcp ``` ## Then, for each person Each person who uses the agent signs in with their own Sleuth account the first time, then asks questions in Copilot or Teams. Their projects are their own. On a Team plan, their questions draw on the team's shared credits. ## A checklist for IT - Allow custom connectors, or MCP tools, in Copilot Studio. Your data policies in the Power Platform admin center control this. - Allow the agent to be published to Microsoft 365 Copilot or Teams. - Allow these addresses: `mcp.sleuthapi.com`, `auth.sleuth.studio` (sign-in), `app.sleuth.studio`, `public-api.sleuth.studio`, and `storage.googleapis.com`. [Your data in AI tools](ai-tools-data.md) Sleuth isn't in Microsoft's marketplace, so this is the way to add it to Copilot. [Security overview](security.md) has the answers IT usually asks for. ## What it costs Sleuth charges nothing extra for Copilot: a cell uses the same credits as in Studio. Microsoft's own Copilot Studio licensing is between your company and Microsoft. ## What you'll see in Copilot Sleuth's answers, with links. Open a project link to see the full datasets, charts, and sources in Studio. ## Related - [Fix connection problems](ai-troubleshooting.md): The most common problems connecting Sleuth to an AI tool, and how to fix each one. - [Security overview](security.md): The short answers for your security team, and where to find the evidence. - [Prompts to start with](ai-prompts.md): Fill in the brackets and paste a prompt into your AI tool. Sleuth shows you its plan before it builds anything. *Updated November 9, 2026.* Source: md/setup-copilot.md --- # Set up Sleuth in other AI tools Cursor, Gemini, VS Code, and agents your team builds connect the same way. [See the MCP reference](mcp-overview.md) Any AI tool that supports remote MCP servers can connect to Sleuth. ## The general steps 1. **Add a remote MCP server**, sometimes called a custom connector, in your tool's settings. 2. **Paste the server URL.** The connection uses streamable HTTP. 3. **Sign in** with your Sleuth account when your tool asks. Sign-in uses OAuth, so you don't need an API key. ``` https://mcp.sleuthapi.com/mcp ``` ## Cursor [Add Sleuth to Cursor](cursor://anysphere.cursor-deeplink/mcp/install?name=sleuth&config=eyJ1cmwiOiJodHRwczovL21jcC5zbGV1dGhhcGkuY29tL21jcCJ9), or add an MCP server with the server URL in Cursor's MCP settings. Cursor asks you to sign in. ## Tools with a config file Add Sleuth to the tool's MCP settings file, then sign in when prompted: ``` { "mcpServers": { "sleuth": { "url": "https://mcp.sleuthapi.com/mcp" } } } ``` ## Agents your team builds Agents that a person signs in to connect over MCP the same way as Claude does, on any plan. Jobs that run without a person, like a nightly refresh, need an API key, which comes with Enterprise. The [MCP reference](mcp-overview.md) lists every tool and how a research run works, and [API](api.md) covers keys. ## Related - [Sleuth MCP server](mcp-overview.md): Connect any MCP client or agent to Sleuth: run research, look up the knowledge base, build Explore datasets, and search the research library, with 52 tools. - [API](api.md): Sleuth’s REST API lets your own software run Sleuth research without the Studio screen. It comes with Enterprise and powers the MCP server. - [Fix connection problems](ai-troubleshooting.md): The most common problems connecting Sleuth to an AI tool, and how to fix each one. *Updated November 9, 2026.* Source: md/setup-other.md --- # Prompts to start with Fill in the brackets and paste a prompt into your AI tool. Sleuth shows you its plan before it builds anything. [Set up your AI tool](ai-tools-overview.md) Each prompt asks for a lighter version of one of Sleuth's workflows. Put in what you know offhand, and keep the rest of the wording: it tells Sleuth what to build. For the full workflow, with its criteria and steps, use Studio. **Indication selection** ```text Our lead asset is a [modality] against [target]. Use Sleuth to rank the indications we could develop it in on unmet need, competition, and trial size, and show why each one ranks where it does. ``` **Asset scouting** ```text Use Sleuth to find unpartnered [modality] assets in [indication] at Phase 1 or later. Show the owner, the stage, the latest data, and any deal history for each, and rank them on how well they fit a mid-sized pharma's pipeline. ``` **Pharma partnering** ```text We're looking for a partner for our [modality] in [therapeutic area]. Use Sleuth to find which of the top 20 pharma are thin there. For each, show what they've bought or said publicly, and how our asset fits. ``` **Asset assessment** ```text We're weighing a deal for [asset name], a [modality] in [indication]. Use Sleuth to pull comparable licensing deals from the last five years, with upfronts and royalty terms, and set them against sell-side peak sales estimates. ``` **Graveyard analysis** ```text Use Sleuth to list every [target] program that has stopped in [indication], with the stage it reached and the reason given, and what the programs still running do differently. ``` **Standard of care evolution** ```text Our asset is a [modality] for [indication]. Use Sleuth to map today's standard of care by line of therapy, list the late-stage readouts that could change it in the next three years, and say which line our first pivotal trial should enroll. ``` ## Tips - **Say "use Sleuth"** so your AI tool calls Sleuth instead of searching the web. - **Approve or edit the plan** when Sleuth shows it. That's the moment to narrow the scope. - **Ask follow-ups** in the same chat. Questions about results Sleuth already built don't use credits. - **Open the project in Studio** to check values cell by cell or share it with your team. ## Related - [Workflows](workflows-overview.md): Six guided analyses for the decisions biopharma teams make most, each run on your own program, brief, or asset in four steps. - [Set up Sleuth in Claude](setup-claude.md): Connect Sleuth on claude.ai, Claude Desktop, and Claude Code. - [How credits work](how-credits-work.md): What a credit buys, what is free, and how to tell what a job will use before it runs. *Updated November 9, 2026.* Source: md/ai-prompts.md --- # Fix connection problems The most common problems connecting Sleuth to an AI tool, and how to fix each one. [Email support](get-help.md) ### Opening the server URL in a browser shows "Method not allowed." That's expected. The URL is for your AI tool to connect to, and a browser can't open it. ### I signed in, but Sleuth says I don't have access. Your account may still be waiting for approval. Once the email saying you're in arrives, sign in again from your AI tool. If you connected before you had an account, your AI tool can request access for you: ask it to get you set up with Sleuth. ### My AI tool answered without using Sleuth. Check that Sleuth is turned on for that chat: in Claude, it's in the tools menu of the message box. Or start your message with "Use Sleuth to". [Set up Sleuth in Claude](setup-claude.md) ### I added Sleuth, but nothing happens. Restart your AI tool: a connector added mid-session may not load until you do. Then check that you finished the sign-in step, which is separate from adding the connector. ### Studio says I'm not connected, but Claude can use Sleuth. You're connected. Studio's setup check can miss a working connection. If Claude can list your Sleuth projects, it's working. ### I joined a team workspace, and now my AI tool can't reach my work. Disconnect Sleuth in your AI tool and connect again, so it signs in to your new workspace. ### I added Sleuth in Copilot Studio, but Copilot doesn't use it. The tool belongs to the agent you added it to. Publish that agent to Microsoft 365 Copilot or Teams, and ask your question there. ### Copilot Studio won't let me add Sleuth, or sign-in is blocked. Your company's policies decide this. Ask IT to allow custom connectors or MCP tools in Copilot Studio, publishing to Microsoft 365 Copilot or Teams, and Sleuth's addresses. [A checklist for IT](setup-copilot.md) ### A request is taking a while. Larger datasets take longer to build. Your AI tool checks back on the project, and you can watch it build in Sleuth Studio. ### It stopped and asked me to approve a plan. Research runs wait for your approval unless auto-approve is on. Approve in the chat, or turn on auto-approve in Studio settings to skip it. [Plans and approvals](plans-approvals.md) ### It says I'm out of credits. Your AI tool shows Sleuth's message as it is, with a link to your plan. The fix is the same as in Studio. [If you run out of credits](out-of-credits.md) ### A download link doesn't work. Download links are short-lived. Ask your AI tool for a new one. If Claude says it can't reach the storage host, that's Claude's own sandbox, not Sleuth: open the link yourself in your browser. ### I can't download a broker report. Sleuth returns the relevant passages and page images from broker research, not the original PDF. [Sell-side research](sell-side.md) ## Still stuck? Email support@sleuthinsights.com with the AI tool you're using and what you see, and we'll help you connect. ## Related - [Set up Sleuth in Claude](setup-claude.md): Connect Sleuth on claude.ai, Claude Desktop, and Claude Code. - [Set up Sleuth in Microsoft Copilot](setup-copilot.md): Add Sleuth to an agent in Copilot Studio, then use it in Microsoft 365 Copilot or Teams. - [Get help](get-help.md): Where to ask, depending on what you need. *Updated November 9, 2026.* Source: md/ai-troubleshooting.md --- # Get help Where to ask, depending on what you need. [Open Studio and chat with us](https://app.sleuth.studio) | You need | Go to | | --- | --- | | Help with Studio, a project, or your account | Press the chat button in Sleuth Studio to reach Sleuth's team, or email support@sleuthinsights.com | | Help connecting an AI tool | [Fix connection problems](ai-troubleshooting.md), then support@sleuthinsights.com | | A problem with a charge, a card, or a refund | support@sleuthinsights.com | | A value you think is wrong | Flag the cell in Review mode, or email support@sleuthinsights.com with the project link and the cell | | A quote, invoicing, a purchase order, vendor forms like a W-9, or Enterprise | sales@sleuthinsights.com or [talk to us](https://www.sleuthinsights.com/talk-to-us) | | Your security team's questions, the SOC 2 report, or the subprocessor list | security@sleuthinsights.com, and the [Trust Center](https://trust.sleuthinsights.com) | | To have your personal data deleted | privacy@sleuthinsights.com | | To build your first project with someone | Book 30 minutes with Sleuth's team. [Talk to us](https://www.sleuthinsights.com/talk-to-us) | | To trade notes with other strategy and BD people | The Sleuth community on Slack. [Join the community](https://www.sleuthinsights.com/community) | | Sleuth's analysts to do the work | [Concierge and custom datasets](concierge-custom-datasets.md) | ## Help us help you When you write in about a project, include the project's name or link, what you asked, and what you expected to see. For a billing question, include your workspace's name. Email to support opens a ticket, and Sleuth's team replies by email. ## If Sleuth is down A banner in Studio says so while Sleuth fixes it. Your AI tool can check too: ask it to check Sleuth's service health. ## Can't find it here? Press the chat button in Studio, and a person on Sleuth's team will help. ## Related - [Frequently asked questions](faq.md): Short answers to what teams ask before and after they start. - [Fix connection problems](ai-troubleshooting.md): The most common problems connecting Sleuth to an AI tool, and how to fix each one. - [Concierge and custom datasets](concierge-custom-datasets.md): When you want Sleuth's analysts to do the work, or a dataset built to your spec and kept current. *Updated November 9, 2026.* Source: md/get-help.md --- # Frequently asked questions Short answers to what teams ask before and after they start. [Start free](https://app.sleuth.studio/signup) ## About Sleuth ### What is Sleuth Insights? Sleuth Insights is AI software for biopharma strategy, business development, and competitive intelligence teams. Sleuth builds analysis for questions like which indication to lead with or which pharma to pitch. Every plan includes three parts: analysis built for your question, an explorable knowledge base, and searchable documents. [Welcome to Sleuth](welcome.md) ### How is Sleuth different from ChatGPT or Claude? Sleuth gives ChatGPT and Claude the library they're missing: a biopharma knowledge base, already matched and checked, with a source and a confidence score on every value. On their own, they rebuild each landscape from web searches, can't reach sell-side research or China's registries in Chinese, and get some of it wrong without telling you where. Connect Sleuth, and Claude, ChatGPT, or Copilot work from the same knowledge base as Studio. [Sleuth in AI tools](ai-tools-overview.md) ### Can Sleuth replace Evaluate, Cortellis, Citeline, or GlobalData? For most strategy work, yes. Every plan includes the same kinds of records: programs by target, modality, and phase, with catalysts and sales forecasts; trials from registries worldwide; and deals with terms where they're disclosed. Sleuth adds sell-side research and China coverage, and browsing never uses credits. [What's in the knowledge base](knowledge-base.md) ### Who uses Sleuth? Business development, strategy, and competitive intelligence teams at biotech and pharma companies, plus the investors and bankers who work with them, from nine-person biotechs to several teams inside a top-10 pharma. ### How long does an analysis take? Longer than a chat answer, because Sleuth plans the project, builds and checks a dataset, then builds the analysis. Larger landscapes take longer. You can watch it build, or leave and get an email when it's done. ## Data ### Where does Sleuth's data come from? From the organizations that publish it: trial registries, regulators, patent offices, journals and medical meetings, companies, news, and sell-side research from 35+ providers, including Chinese sources read in Chinese. [Every source](sources.md) ### Can it see paywalled journals? Sleuth reads published papers and preprints that are open access or published under terms that allow indexing, plus abstracts and metadata for the rest of the literature, and conference abstracts and posters. Sell-side research, which usually sits behind a paywall, is included in every plan. [Every source](sources.md) ### How current is the data? Sleuth updates over 95% of its sources in real time. Turn on monitoring to keep your own datasets current as new data lands. [How fresh the data is](data-freshness.md) ### Does Sleuth cover preclinical programs and private companies? Yes. About 60% of the 150K+ drugs in the knowledge base are preclinical, including programs that so far appear only in patents, and private companies come with their pipelines, deals, and financings. ### Does Sleuth cover Chinese programs? Over 22,000 of them, from China's registries, regulators, exchanges, company posts, and broker research, translated and matched to the same drug records as everything else. [China coverage](china.md) ### Which banks' research is included? Research from 35+ providers: global investment banks, US healthcare and mid-market banks, European brokers, independent and sponsored research, consulting and industry research, and historical archives. [Ask us](https://www.sleuthinsights.com/talk-to-us) for the full list. [Sell-side research](sell-side.md) ## Accuracy ### How accurate is Sleuth? Every value links to its source and has a confidence score, based on how good its best source is, whether independent sources agree, and how recent it is. When Sleuth can't find a source, it marks the value unknown instead of filling it in. [Confidence scores](confidence.md) ### Do I have to check every cell? No. Values with strong sources that agree score high, and Review mode shows you only the cells worth a second look, with the reason. [Review mode](review-mode.md) ### How does Sleuth avoid hallucinations? Sleuth checks every value before you see it: it matches and normalizes every record, runs AI checks at each step, flags disagreements between sources instead of smoothing them over, and labels every value sourced, inferred, or unknown, so you can check any number in one click. ## Pricing and credits ### How much does Sleuth cost? $249 a month for Starter, for one person with 3,000 credits, and from $499 a month for Team, with 6,000 credits shared by any number of people. Start with a 14-day free trial that needs no card. [Plans and pricing](plans.md) ### Is it priced per seat? No. Team has no per-seat price: any number of people share one pool of credits. Starter is a one-person plan. [Plans and pricing](plans.md) ### What's a credit? One credit is one analytical cell: a judgment Sleuth makes about one item in your research. Scoring 150 assets on 3 criteria uses 450 credits. Browsing the knowledge base, searching documents, and asking about your results never use credits. [How credits work](how-credits-work.md) ### How many credits will we need? Count the analytical cells in the work you do each month: rows times the judgment columns. A team scoring two landscapes of 150 assets on 3 criteria each month uses about 900 credits, before follow-ups that build new columns. Sleuth shows you its plan before each run, so you can see the size of the job first. ### What happens if we run out? Runs already going finish. New credit actions pause until you upgrade or your next credits arrive, and everything else keeps working. [If you run out of credits](out-of-credits.md) ### Do unused credits roll over? Yes. Monthly plans bank up to two months of credits; annual plans carry up to two extra months into the next year. Bonus and trial credits never expire. [Rollover and bonuses](rollover-bonuses.md) ### Can we pay annually or by invoice? Annual Team plans cost 10 times the monthly price, so you get two months free, and the credits arrive up front. Invoices are available on annual Team plans from 10,000 credits a month, and on Enterprise. For a purchase order or vendor forms like a W-9, email sales@sleuthinsights.com. ### Can we cancel or change plans? Yes, any time, with no cancellation fee. Upgrades take effect right away; downgrades and cancellations take effect at the end of the period you paid for. [Change your plan](change-plan.md) ### Is there a free plan? No, but there are two free ways in: the 14-day free trial, with no card, and, if your company has raised under $10 million, a free year for your whole team with Sleuth for Startups. [Sleuth for Startups](startups.md) ### Can I get Sleuth for one project? Plans are monthly, so you can take Starter or Team for the months a project runs and cancel after. For a project with a fixed scope, Sleuth's analysts can do the work with Concierge. [Concierge](concierge-custom-datasets.md) ## AI tools ### Can we use Sleuth in Claude, ChatGPT, or Copilot? Yes, on every plan. Add Sleuth to the AI tool your company already uses, sign in with your Sleuth account, and ask there. [Sleuth in AI tools](ai-tools-overview.md) ### Does it cost extra? No. Connecting is included in every plan, and a cell costs the same credits there as in Studio. ### Does Sleuth have an MCP server? Yes, at https://mcp.sleuthapi.com/mcp, so anything that supports MCP connects to Sleuth on any plan, including Gemini and agents your team builds. Agents that run without a person signing in need an API key, which comes with Enterprise. [MCP reference](mcp-overview.md) ### Does Sleuth have an API? Sleuth's API comes with Enterprise, for applications that call Sleuth directly. Agents that support MCP don't need it. [API](api.md) ## Your data ### Is our data private? Yes. On every plan, your questions, uploads, and results stay in your workspace and are never shown to other customers. [How your data is used](data-use.md) ### Is Sleuth SOC 2 compliant? Yes. Sleuth is audited under SOC 2 Type II, and your data is encrypted in transit and at rest and stored in the United States. Email security@sleuthinsights.com for the report. [Security overview](security.md) ### Can we use our own documents? Yes, on any plan. Upload files and Sleuth reads them alongside everything else, or start a project from your own spreadsheet. [Use your own documents and data](own-data.md) ### What can we export? PowerPoint decks, reports, charts, and every dataset as a download, or the whole project as one file to hand to Claude or your team. [Outputs and exports](outputs.md) ## Getting started ### How do I get started? Press Start free, sign up, and choose Studio, your AI tool, or both. Sleuth's team approves new accounts, and you'll get an email when you're in. [Sign up and get access](sign-up.md) ### Can my team try it together? The free trial is for one person. To try Sleuth as a team, choose a Team plan, which you can cancel any time, or talk to us about a team trial. ### Can I sign up with Gmail? Yes. With a personal email address, Sleuth asks for your LinkedIn profile instead of your company's website, and Sleuth's team approves the account. [Sign up and get access](sign-up.md) ### Is there a desktop app? No. Studio runs in your browser, on computers, tablets, and phones. To use Sleuth from a desktop app, connect it to Claude Desktop. [Set up Sleuth in Claude](setup-claude.md) ### Do teammates see each other's projects? Each project belongs to the person who made it. Send a colleague a copy to build on, or transfer it to them. [Share your work](share-work.md) ### Can I give Sleuth's outputs to my clients? Plans cover your own team's work, and Sleuth's terms don't allow reselling its content. If you're delivering Sleuth's analysis to clients, [talk to us](https://www.sleuthinsights.com/talk-to-us) so your plan covers it. ### What should I run first? Start from a workflow built for a common job: indication selection, asset scouting, pharma partnering, asset assessment, graveyard analysis, or standard of care evolution. [Workflows](workflows-overview.md) ## Related - [Get help](get-help.md): Where to ask, depending on what you need. - [Glossary](glossary.md): The words you'll see in Sleuth and in these docs. - [Plans and pricing](plans.md): Starter for one person, Team for any number of people, Enterprise for company-wide use, and what a credit costs. *Updated November 9, 2026.* Source: md/faq.md --- # Glossary The words you'll see in Sleuth and in these docs. [Open Studio](https://app.sleuth.studio) **Analytical cell**: A cell where Sleuth makes a judgment for your question, like an asset's fit with your strategy. Uses one credit. [How credits work](how-credits-work.md) **Analyze**: The part of Studio where you ask questions in your own words. Each question becomes a project. [Ask a question](ask-question.md) **Auto-approve**: A setting that lets plans, dataset designs, and dataset changes go ahead without your click. [Plans and approvals](plans-approvals.md) **Confidence score**: A 0 to 100 score on every value, from its source quality, agreement between sources, and recency. [Confidence scores](confidence.md) **Credit**: What Sleuth's analysis is paid in. One credit is one analytical cell. **Custom column**: A column you define that Sleuth fills in for every row. [Add your own columns](custom-columns.md) **Dataset**: A table Sleuth builds for your question, with a source and a score on every value. [Datasets](datasets.md) **Dataset design**: The columns a dataset will have, which you approve before Sleuth builds it. **Document search**: Searching and reading the documents behind the knowledge base. [Search documents](document-search.md) **Explore**: The part of Studio where you filter the knowledge base and build datasets yourself, without credits. [Explore](explore.md) **Fact cell**: A cell whose value the knowledge base already has, like phase or sponsor. Free. **Initiative**: A folder for related projects. [Projects and initiatives](share-work.md) **Knowledge base**: All of Sleuth's data, connected: drugs, trials, companies, targets, deals, and the documents behind them. [What's in the knowledge base](knowledge-base.md) **MCP**: Model Context Protocol, the standard way AI tools like Claude connect to outside data. Sleuth's MCP server is how they connect to Sleuth. [MCP reference](mcp-overview.md) **Monitoring**: Keeping a dataset current as new data lands, with versions and alerts. [Monitoring](monitoring.md) **Plan**: The steps Sleuth proposes before it builds anything. Also the subscription you pay for, like Starter or Team. **Project**: Where a question or workflow run lives, with its chat, datasets, files, and alerts. **Review mode**: A view of a dataset that shows only the cells worth checking. [Review mode](review-mode.md) **Sell-side research**: Reports from banks and brokers, included in every plan. [Sell-side research](sell-side.md) **Sourced, inferred, unknown**: How Sleuth labels a value: stated by a linked document, worked out from sources, or not publicly available. **Workflow**: A multi-step job Sleuth knows how to run, like indication selection, with inputs, a plan, and outputs. [Workflows](workflows-overview.md) **Workspace**: Your account's space in Sleuth. Personal on the trial and Starter, shared on Team. **Your context**: What Sleuth keeps about you and your company so you don't repeat it. [Your context](your-context.md) ## Related - [Frequently asked questions](faq.md): Short answers to what teams ask before and after they start. - [Welcome to Sleuth](welcome.md): What Sleuth does, where you use it, and where to go in these docs for each job. *Updated November 9, 2026.* Source: md/glossary.md --- # Concierge and custom datasets When you want Sleuth's analysts to do the work, or a dataset built to your spec and kept current. [Talk to us](https://www.sleuthinsights.com/talk-to-us) ## Concierge Sleuth's analysts take your question and deliver a finished report or deck, and the project behind it is saved to your Studio account so you can check every source. They start the day you send it, and you see the first findings within days. - **Good for:** a board meeting, a partnering conference, a deal review, or anything you need done well and don't have the hours for. - **How it's priced:** Sleuth quotes each question, for a fraction of what a consulting project costs. Start with one project, or keep the analysts on call and send questions as they come up. - **How it's checked:** an analyst checks every Concierge delivery before it reaches you. ## Custom datasets A dataset Sleuth builds to your spec and keeps current, like a competitor tracker or a target list your team updates by hand today. You set the rows, the columns, and the rules, and Sleuth delivers it on your schedule, into Studio, a spreadsheet, your data warehouse, or your CRM. Quality checks built for each dataset run on every delivery. ## Get started [Talk to us](https://www.sleuthinsights.com/talk-to-us) about your question or dataset, and we'll scope it with you. ## Related - [Enterprise](enterprise.md): What Enterprise adds for companies rolling Sleuth out across teams and systems. - [Get help](get-help.md): Where to ask, depending on what you need. - [Keep a project current with monitoring](monitoring.md): Turn on monitoring and Sleuth updates your datasets as new trials, filings, and deals come in. *Updated November 9, 2026.* Source: md/concierge-custom-datasets.md --- # Workflows Six guided analyses for the decisions biopharma teams make most, each run on your own program, brief, or asset in four steps. [Open the workflow library](https://app.sleuth.studio) A workflow is an analysis Sleuth has already planned for a common job, like choosing a lead indication or deciding whether to license an asset. You describe your case, choose what counts, check the plan, and Sleuth builds the analysis. Every value in it has a source and a confidence score, and the datasets stay in your project to keep working in. Home suggests workflows based on what you told Sleuth at onboarding, and **Workflows** in the Studio sidebar opens the full library. ## The six workflows | Workflow | The question it answers | What you get | | --- | --- | --- | | [**Indication selection**](wf-indication-selection.md) | Which indication should we lead with? | A recommendation on the lead, the second, and the ones to drop, with the evidence behind every disease | | [**Graveyard analysis**](wf-graveyard.md) | Why did drugs like ours fail? | One main cause for each discontinued drug like yours, and whether your design fixes it | | [**Standard of care evolution**](wf-standard-of-care.md) | Who should our first pivotal trial enroll? | The market you'll launch into, redrawn for each readout, and the patient groups ranked | | [**Pharma partnering**](wf-pharma-partnering.md) | Which pharma should we pitch? | A ranked list of pharma, with what to lead with and what to bring to each | | [**Asset scouting**](wf-asset-scouting.md) | Which asset should we license first? | A ranked shortlist scored on your criteria, with who holds each asset's rights | | [**Asset assessment**](wf-deal-assessment.md) | Should we license this asset? | A sales forecast built input by input, a score, and a recommendation on terms | The first three are about your own program. The last three are about other companies and their assets. In Studio, workflows carry their own names (Novel Platform to Positioning, for example). Each page here gives both. ## How every workflow runs 1. **Describe your case.** Say what your platform, program, or asset does, the decision you face, and anything you already know. Add a deck or a document if you have one. 2. **Decide what counts.** Choose the criteria that matter, like clinical validation or competitive room, and how much each one counts. 3. **Shape the plan.** Sleuth shows the steps it will run, the outputs, and your credit balance. Add or drop scope, choose **Pause at each step** or **Run end to end**, and press **Start**. [Plans and approvals](plans-approvals.md) 4. **Get the analysis.** A recommendation with the evidence behind it, the datasets to keep working in, and outputs you can export. [Outputs and exports](outputs.md) > **Credits:** The plan shows the size of the job before anything runs: rows times analytical columns is the credits it will use, one credit per analytical cell. Questions about results Sleuth already built don't use credits. [How credits work](how-credits-work.md) ## Choosing between them | If you're deciding | Run | | --- | --- | | Where a new platform or asset should go first | [Indication selection](wf-indication-selection.md) | | Where to take a program next without repeating what ended the drugs before it | [Graveyard analysis](wf-graveyard.md) | | Which patients a first pivotal trial should enroll | [Standard of care evolution](wf-standard-of-care.md) | | Which pharma to approach about your platform, and with what | [Pharma partnering](wf-pharma-partnering.md) | | Which outside asset to license against a brief | [Asset scouting](wf-asset-scouting.md) | | Whether one specific asset is worth licensing, and on what terms | [Asset assessment](wf-deal-assessment.md) | | Anything else, like mapping a target, a disease, or a modality | [Ask your own question](wf-own-question.md) | Indication selection and graveyard analysis both ask where a platform goes next. Indication selection finds where you'd be first. Graveyard analysis finds where you'd fail the way others did, and many teams run both. ## Workflows in Claude, ChatGPT, and Copilot Workflows run in Studio. From an AI tool, start with the prompt written for the same job: it asks Sleuth for a lighter version of the same analysis in plain words. [Prompts to start with](ai-prompts.md) ## Related - [Run your first analysis](first-analysis.md): From sign-in to a finished analysis you can check and export, in five steps. - [Plans and approvals](plans-approvals.md): Sleuth shows you what it will build before it builds it. Here is what you approve and how to skip approvals. - [Prompts to start with](ai-prompts.md): Fill in the brackets and paste a prompt into your AI tool. Sleuth shows you its plan before it builds anything. *Updated November 9, 2026.* Source: md/workflows-overview.md --- # Indication selection Choose the lead indication for your platform: Sleuth tests each disease it could treat for proof your mechanism helps, maps who’s already there, and sizes the patients. [Run Indication selection](https://app.sleuth.studio) > **Same workflow, two names:** In Studio's workflow library, Indication selection is called **Novel Platform to Positioning**. It's the same workflow as the [Indication Selection](https://www.sleuthinsights.com/use-cases/indication-selection) page on sleuthinsights.com. ## Use it when - Your platform could treat many diseases and you can fund one program first. - You need a defensible lead before a raise, a board meeting, or a partnering push. - You're the CEO, CSO, or first BD hire at a platform biotech, or you're weighing a lead asset's next indication. ## What to have ready How your platform works, what sets it apart, the decision you face, and any diseases you're already weighing. Add your deck if you have one. You can start from a single lead asset instead of a platform: describe it the same way. ## How it runs 1. **Describe your platform.** The mechanism, what it does better, and the areas you'd consider. 2. **Decide what counts.** Choose what matters most, such as clinical validation, unmet need, competitive room, and fit with your platform, and how much each one counts. 3. **Shape the plan.** Sleuth shows its three steps: test every candidate disease, map the competition, and size each position. Add a disease or drop one before it runs. 4. **Get the analysis.** A recommendation on where to lead, with the evidence behind every disease, and the datasets to keep working in. ## What you get | Your question | What Sleuth gives you | | --- | --- | | Does our mechanism work in this disease? | Every disease your platform could treat, tested for human proof that your mechanism helps and for a fit with how you'd dose it. A failed trial counts against a disease, and Sleuth records why it ruled out each one. | | Who's already there, and how close are they to us? | Every program in each shortlisted disease, from preclinical to approved, sorted by how close its mechanism comes to yours. | | How many patients, and what are they worth? | The patients each position would serve in your markets, and what they're worth at branded prices against what's spent on them today. | | Where should we lead, and what must the first trial show? | A recommendation on the lead, the second, and the ones to drop, with what your first trial has to show in each. | ## A finished example **Where should an antibody-removal platform lead? Sleuth says AQP4-positive NMOSD.** Sleuth tested 9 candidate diseases, shortlisted 3, and found that none of NMOSD's 94 programs removes the antibody that causes it. [See the full example](https://www.sleuthinsights.com/use-cases/indication-selection) ## Questions ### What if we already have diseases in mind? Add them in step 1. Sleuth tests them alongside every other candidate, so you can see how your list compares. ### Why was a disease ruled out? Sleuth lists the test each one missed and the reason, such as no single antibody that causes the disease, or a removal trial that missed its endpoint. ### How many diseases does it look at? Every disease your platform could treat in the areas you name, not only the ones on your list. In the example, Sleuth tested 9 candidates and shortlisted 3. ### Does it size the market? Yes. For each shortlisted disease, Sleuth sizes the patients in your markets and what they'd be worth at branded prices, next to what's spent on them today. ### Can I change the weights after it runs? Yes. Change a weight, and Sleuth ranks the diseases again. > **Credits:** The plan shows the size of the job before anything runs: rows times analytical columns is the credits it will use. [How credits work](how-credits-work.md) ## Next questions on the same project [Graveyard analysis](wf-graveyard.md) checks whether your lead repeats an old failure. [Standard of care evolution](wf-standard-of-care.md) picks the patients for the first pivotal trial. [Pharma partnering](wf-pharma-partnering.md) lines up who to pitch. ## Related - [Graveyard analysis](wf-graveyard.md): Find out why drugs like yours failed: for every one that was discontinued, Sleuth works out why it ended and whether your design fixes that failure. - [Standard of care evolution](wf-standard-of-care.md): Pick the patients for your first pivotal trial: Sleuth maps today’s standard of care line by line, dates every readout that could change it before you launch, and redraws the market for each result. - [Pharma partnering](wf-pharma-partnering.md): Know which pharma to approach, and with what: Sleuth maps the top 25 pharma’s pipelines by franchise, matches your best program to where each one is thin, and lists the proof each will ask for. *Updated November 9, 2026.* Source: md/wf-indication-selection.md --- # Graveyard analysis Find out why drugs like yours failed: for every one that was discontinued, Sleuth works out why it ended and whether your design fixes that failure. [Run Graveyard analysis](https://app.sleuth.studio) *In Studio: **Graveyard Analysis*** ## Use it when - You're choosing where to take a program or platform next: an organ, an indication, a target, or a modality. - You need to show a board or partner that you won't repeat what ended the drugs before yours. ## What to have ready What your program or platform does, the decision you face, and the options you're weighing, such as which organ, indication, or target to go after next. The options can also be assets you'd license in. ## How it runs 1. **Describe your program.** What it does and the options in front of you. 2. **Decide what counts.** Choose how to weigh each option: how often its drugs failed on their own data, whether your design fixes that failure, and how early a first trial could show it. 3. **Shape the plan.** Sleuth shows its three steps: find every discontinued drug like yours, find out why each one ended, and compare them with yours. Widen what counts as like yours before it runs. 4. **Get the analysis.** A priority for each option, what your first trial has to show, and a table of every discontinued drug with its cause. ## What you get | Your question | What Sleuth gives you | | --- | --- | | Which drugs like ours were discontinued? | Every drug like yours that entered the clinic, each counted once with its status checked, and for each discontinued one, the stage it reached. | | Why did each one end? | One main cause for each, from the sponsor's statement or its last reported data, so failures on a drug's own data stay apart from portfolio and financing decisions. | | Does our design fix that failure? | Each option you're weighing, set against how its drugs failed and marked as fixing that failure or exposed to it, with what your first trial would have to show. | ## A finished example **Where should a liver siRNA company go next? Sleuth says skeletal muscle.** 233 discontinued RNA drugs, sorted by organ and cause. Every muscle failure on data was a delivery shortfall, the problem the company already solves in the liver. [See the full example](https://www.sleuthinsights.com/use-cases/graveyard-analysis) ## Questions ### Is there a reason for every discontinued drug? Where the sponsor gave one, Sleuth uses it. Where it didn't, Sleuth infers the cause from the last reported data and labels it inferred. ### Does it separate business decisions from failures? Yes. Each drug gets one main cause, so a portfolio or financing decision stays apart from a failure on the drug's own data. In the example, portfolio decisions were the most common single cause. ### Can the options be targets or modalities? Yes. The options can be indications, organs, targets, modalities, or assets you'd license in. ### Does it cover drugs still in development? Yes. Sleuth compares your program with the drugs still in development too, so you can see who's already working on the failure you'd fix. ### What does it say my first trial has to show? For each option, Sleuth states the requirement. In the example, that's drug levels and knockdown in muscle biopsies from the first patients, at a tolerable dose. > **Credits:** The plan shows the size of the job before anything runs: rows times analytical columns is the credits it will use. [How credits work](how-credits-work.md) ## Next questions on the same project [Indication selection](wf-indication-selection.md) finds where you'd be first. [Asset scouting](wf-asset-scouting.md) finds an asset to bring in. [Standard of care evolution](wf-standard-of-care.md) picks the patients for a first pivotal trial. ## Related - [Indication selection](wf-indication-selection.md): Choose the lead indication for your platform: Sleuth tests each disease it could treat for proof your mechanism helps, maps who’s already there, and sizes the patients. - [Asset scouting](wf-asset-scouting.md): Find the asset worth licensing first: Sleuth scores every asset that fits your brief on your own criteria and checks who still holds the rights. - [Standard of care evolution](wf-standard-of-care.md): Pick the patients for your first pivotal trial: Sleuth maps today’s standard of care line by line, dates every readout that could change it before you launch, and redraws the market for each result. *Updated November 9, 2026.* Source: md/wf-graveyard.md --- # Standard of care evolution Pick the patients for your first pivotal trial: Sleuth maps today’s standard of care line by line, dates every readout that could change it before you launch, and redraws the market for each result. [Run Standard of care evolution](https://app.sleuth.studio) *In Studio: **Standard of Care Evolution*** ## Use it when - You lead clinical strategy, or you're the CEO, with an early asset and a first pivotal trial to design. - You need to know which line of therapy and which patients give you the best start, given what could change before launch. - You're building a forecast and need patients and time on treatment by line in your launch year. ## What to have ready What your asset has shown, the edge you expect, the decision you face, and the markets and year you expect to launch. ## How it runs 1. **Describe your asset.** Its data so far, its edge, and your launch markets and year. 2. **Decide what counts.** Choose how to rank each patient group, such as the result to beat, patient numbers, time to readout, and competition, and how much each one counts. 3. **Shape the plan.** Sleuth shows its three steps: map today's standard of care, find what could change it, and project where your asset should start. Add a trial to watch before it runs. 4. **Get the analysis.** The patient groups ranked for your first pivotal trial, the readouts that could change the order, and the market map as a dataset. ## What you get | Your question | What Sleuth gives you | | --- | --- | | What's the standard today, and the bar in each line? | Each line of therapy and patient group, with its standard regimen, the result to beat, how many patients there are, and how long they stay on treatment. | | What could change it before we launch? | Every late-stage trial that could change a standard by your launch year, with its comparator, its eligibility, and a dated readout, including your own next data. | | Where should we start if each readout goes either way? | The market redrawn for each result, and the patient groups ranked for your first pivotal trial, with what would change the order. | ## A finished example **Where should a CDK2 inhibitor's first pivotal trial start? After a CDK4/6 inhibitor.** The HR+/HER2-negative breast cancer market of 2031, redrawn for 25 readouts. Second line stays first unless the 2028 class readouts or the drug's Phase 1b fail. [See the full example](https://www.sleuthinsights.com/use-cases/standard-of-care) ## Questions ### Can I use it for a forecast? Yes. The map sizes the patients and the time on treatment in each line for your launch year, which is where a forecast starts. ### How are readouts dated? From company guidance or the FDA where there is one, and from the trial registry or other sources where there isn't. The map marks which is which. ### Does it count our own next readout? Yes. Your next data, such as a Phase 1b, goes on the timeline with every other readout, and the ranking shows where you'd start if it fails. ### Which markets does it cover? The markets you choose. The example maps the U.S. ### Does it cover every line of therapy? Yes. Each line and patient group gets its own standard, benchmark, and patient numbers, including the groups today's pivotal trials left out. > **Credits:** The plan shows the size of the job before anything runs: rows times analytical columns is the credits it will use. [How credits work](how-credits-work.md) ## Next questions on the same project [Pharma partnering](wf-pharma-partnering.md) finds who to partner with for the first line. [Asset assessment](wf-deal-assessment.md) prices an asset against the same market. [Graveyard analysis](wf-graveyard.md) shows what ended the drugs before yours. ## Related - [Pharma partnering](wf-pharma-partnering.md): Know which pharma to approach, and with what: Sleuth maps the top 25 pharma’s pipelines by franchise, matches your best program to where each one is thin, and lists the proof each will ask for. - [Asset assessment](wf-deal-assessment.md): Decide whether to license an asset, and how: Sleuth maps the rivals it will launch against, works out who’s left to treat, benchmarks its data cohort by cohort, and forecasts sales. - [Keep a project current with monitoring](monitoring.md): Turn on monitoring and Sleuth updates your datasets as new trials, filings, and deals come in. *Updated November 9, 2026.* Source: md/wf-standard-of-care.md --- # Pharma partnering Know which pharma to approach, and with what: Sleuth maps the top 25 pharma’s pipelines by franchise, matches your best program to where each one is thin, and lists the proof each will ask for. [Run Pharma partnering](https://app.sleuth.studio) *In Studio: **Platform to Pharma Partnering*** ## Use it when - You're a platform biotech's CEO or BD lead preparing for a partnering conference or a round of outreach. - You want to tailor each pitch to what that pharma is building, instead of sending one deck to everyone. - You need to prepare for one big meeting with the objections and proof in hand. ## What to have ready What your platform has and hasn't shown, the decision you face, such as a conference coming up, and where you want to focus. It works before you have clinical data. ## How it runs 1. **Describe your platform.** Its data so far and the areas you want to focus on. 2. **Decide what counts.** Choose what makes a good partner for you, such as fit with a franchise, deal record, and whether it already owns a competing platform, and how much each one counts. 3. **Shape the plan.** Sleuth shows its three steps: map the pipelines, match your programs and rank the pharma, and prepare each outreach. Add a company or drop one before it runs. 4. **Get the analysis.** A ranked list of pharma, with what to lead with and what to bring for each, and the program matches as a dataset. ## What you get | Your question | What Sleuth gives you | | --- | --- | | What is each pharma building? | Each pharma's programs, deals, and stated priorities, franchise by franchise, with the platforms it already owns or licenses. | | Where does our platform fit them? | What your platform could lead with, matched to each franchise and ranked on cited evidence. Sleuth sets aside any fit that depends on data you don't have yet. | | Who already has something like us? | The pharma that own or license a competing platform, ranked lower with the reason, so you don't pitch a company that's building it in house. | | What will each one ask before a meeting? | For each priority pharma, the case for and against, the objections to expect, the proof to have in hand, and the events to watch before your meeting. | ## A finished example **Which pharma should an immune-cell RNA platform approach first? Sleuth says Biogen.** A fictional platform, matched to the top 25 pharma before a March partnering conference. Given the same question, a general AI model left Biogen out. [See the full example](https://www.sleuthinsights.com/use-cases/pharma-partnering) ## Questions ### Does it work before we have clinical data? Yes. The example platform is preclinical. Sleuth sets aside any fit that depends on data you don't have yet, and lists the proof to get first. ### Can I add a pharma that isn't in the top 25? Yes. Add it when you shape the plan, and Sleuth maps it the same way. ### Can I prepare for one meeting instead of five? Yes. Name the company, and Sleuth builds its case for and against, the objections to expect, and the proof to bring. ### How does Sleuth know what each pharma wants? From each pharma's programs, deals, and stated priorities, read franchise by franchise, with the source behind each one. ### Why is a pharma ranked lower? Sleuth gives the reason beside it, such as owning a competing platform, or a fit that depends on data you don't have yet. > **Credits:** The plan shows the size of the job before anything runs: rows times analytical columns is the credits it will use. [How credits work](how-credits-work.md) ## Next questions on the same project [Indication selection](wf-indication-selection.md) settles what to lead with. [Standard of care evolution](wf-standard-of-care.md) shows the market a partner will weigh. [Monitoring](monitoring.md) keeps each pharma's pipeline current before the meeting. ## Related - [Indication selection](wf-indication-selection.md): Choose the lead indication for your platform: Sleuth tests each disease it could treat for proof your mechanism helps, maps who’s already there, and sizes the patients. - [Asset scouting](wf-asset-scouting.md): Find the asset worth licensing first: Sleuth scores every asset that fits your brief on your own criteria and checks who still holds the rights. - [Keep a project current with monitoring](monitoring.md): Turn on monitoring and Sleuth updates your datasets as new trials, filings, and deals come in. *Updated November 9, 2026.* Source: md/wf-pharma-partnering.md --- # Asset scouting Find the asset worth licensing first: Sleuth scores every asset that fits your brief on your own criteria and checks who still holds the rights. [Run Asset scouting](https://app.sleuth.studio) *In Studio: **Partner Identification*** ## Use it when - You're on a pharma BD or search and evaluation team with a licensing brief. - You need a ranked shortlist your team can defend, with the reason behind every score. - You want to know what's already taken before anyone makes a call. ## What to have ready What you want (the modality, stage, indication, and rights), the decision it feeds, and what matters most to your team. ## How it runs 1. **Describe what you're looking for.** Your brief, in your own words. 2. **Decide what counts.** Pick the criteria, such as white space, target validation, early efficacy and safety, and deal comps, and how much each one counts. 3. **Shape the plan.** Sleuth shows its three steps: build the landscape, define the criteria, and score and rank. Narrow the scope or change a criterion before it runs. 4. **Get the analysis.** A ranked shortlist and a recommendation, the score behind every asset, and the full landscape as a dataset. ## What you get | Your question | What Sleuth gives you | | --- | --- | | What's out there that fits? | Every asset in your scope, at companies of every size and in every region, each counted once with its development stage checked. | | Is it still available? | Each asset's partnering status, checked at the source: unpartnered, licensed out, held with shared rights, or looking for a partner. | | How does it score on what we care about? | Your criteria, in your own words where you want, with set levels behind each score and the fact that decided it. | | Which one do we call first? | A ranked shortlist with a recommendation, what other buyers paid for assets like these, and what to ask each company before talks start. | ## A finished example **Which early T-cell engager should a mid-sized pharma license first? Sleuth says CLN-049.** Sleuth scored 150 early T-cell engagers in oncology on seven criteria and checked who holds the rights. Four of the top ten were already taken. [See the full example](https://www.sleuthinsights.com/use-cases/asset-scouting) ## Questions ### Will it find assets our usual searches miss? Sleuth searches companies of every size, including private companies and programs out of China, and counts each asset once. [China coverage](china.md) ### Can we use our own criteria? Yes. Name them in your own words, set what each score level means, and choose how much each one counts. ### What if an asset hasn't reported data yet? Sleuth leaves that score blank instead of scoring it low, so missing data doesn't push an asset down the list. ### Does it include assets that are already partnered? Yes, marked as licensed out or held with shared rights, so you can see what's taken before you call. In the example, four of the top ten were. ### How early can the assets be? You set the stages in your brief. The example covers first-in-human to Phase 1/2. ### Does it show what other buyers paid? Yes. Deal comps can be one of your criteria, with the deals done on each target and the largest upfront paid. > **Credits:** The plan shows the size of the job before anything runs: rows times analytical columns is the credits it will use. Each asset scored on each criterion is an analytical cell, so a wider scope or more criteria uses more credits. [How credits work](how-credits-work.md) ## Next questions on the same project [Asset assessment](wf-deal-assessment.md) works out what your top pick is worth. [Monitoring](monitoring.md) tells you when a shortlisted asset changes hands or reports data. ## Related - [Asset assessment](wf-deal-assessment.md): Decide whether to license an asset, and how: Sleuth maps the rivals it will launch against, works out who’s left to treat, benchmarks its data cohort by cohort, and forecasts sales. - [Graveyard analysis](wf-graveyard.md): Find out why drugs like yours failed: for every one that was discontinued, Sleuth works out why it ended and whether your design fixes that failure. - [Keep a project current with monitoring](monitoring.md): Turn on monitoring and Sleuth updates your datasets as new trials, filings, and deals come in. *Updated November 9, 2026.* Source: md/wf-asset-scouting.md --- # Asset assessment Decide whether to license an asset, and how: Sleuth maps the rivals it will launch against, works out who’s left to treat, benchmarks its data cohort by cohort, and forecasts sales. [Run Asset assessment](https://app.sleuth.studio) *In Studio: **Deal Assessment from an Asset*** ## Use it when - You're in BD, corporate development, or at a fund, with an asset in front of you and a deal review coming. - You need to check a partner's asking price against what the asset is likely to earn. - You want a portfolio review or partnering story for an asset in your own pipeline. ## What to have ready The asset's name, the patients it's for, what its data shows so far, and the decision you face, such as licensing its U.S. rights. Pick the year you want the forecast for. ## How it runs 1. **Describe the asset.** The asset, its patients, its data, and the deal in question. 2. **Decide what counts.** Choose how to score it: clinical differentiation, competitive room, patients still eligible, commercial size, and development maturity, and how much each one counts. 3. **Shape the plan.** Sleuth shows its four steps: profile the asset and its rivals, map the standard of care in your forecast year, benchmark the data, and forecast and score. Change the forecast year before it runs. 4. **Get the analysis.** A score and a recommendation, the forecast with every input laid out, and the landscape and benchmarks as datasets. ## What you get | Your question | What Sleuth gives you | | --- | --- | | Who will it launch against? | Every rival program in the indication, from Phase 1 to approved, counted once and ordered by stage, so you see where the asset would land in line. | | Who's left to treat by launch? | The standard of care today and in your forecast year, with each trial, label, and guideline change dated, and the share of patients still eligible for the asset. | | Does its data beat theirs? | Efficacy, durability, and safety compared cohort by cohort, with every comparison that isn't like for like flagged. | | What's it worth, and what should we offer? | A sales forecast built input by input, the inputs that move it most, and a recommendation on whether and how to license. | ## A finished example **Should you license enzomenib's U.S. rights? Sleuth says pursue, with conditions.** Sumitomo Pharma's menin inhibitor in relapsed AML, set against every other menin inhibitor in the clinic. Time on drug moves its 2029 sales most. [See the full example](https://www.sleuthinsights.com/use-cases/deal-assessment) ## Questions ### Can we use our own assumptions? Yes. Every forecast input is laid out with its source and range, so you can put in your own price, share, or time on therapy and see what changes. ### What if the trials aren't comparable? Sleuth flags each comparison that isn't like for like, such as pooled doses, a different patient mix, or a combination regimen, beside the number. ### Does it include rivals from China? Yes. They're in the competitive set and marked, as three of enzomenib's are. [China coverage](china.md) ### How far out does the forecast go? To the forecast year you choose. The example forecasts U.S. sales in 2029. ### Does it say what to offer? Yes. The recommendation turns the forecast into terms, such as how large the upfront should be and which conditions to set before exercising an option. ### Can I run it on an asset from our own pipeline? Yes. The same steps work on your own asset, for a partnering story or a portfolio review. > **Credits:** The plan shows the size of the job before anything runs: rows times analytical columns is the credits it will use. [How credits work](how-credits-work.md) ## Next questions on the same project [Asset scouting](wf-asset-scouting.md) finds other assets to compare. [Standard of care evolution](wf-standard-of-care.md) goes deeper on the market the asset launches into. [Graveyard analysis](wf-graveyard.md) shows what ended drugs like it. ## Related - [Asset scouting](wf-asset-scouting.md): Find the asset worth licensing first: Sleuth scores every asset that fits your brief on your own criteria and checks who still holds the rights. - [Standard of care evolution](wf-standard-of-care.md): Pick the patients for your first pivotal trial: Sleuth maps today’s standard of care line by line, dates every readout that could change it before you launch, and redraws the market for each result. - [Add your own columns](custom-columns.md): Ask Sleuth to fill in a column you define, on any dataset in Explore or Analyze. *Updated November 9, 2026.* Source: md/wf-deal-assessment.md --- # When no workflow fits Ask your own question in Analyze. Sleuth plans the work the same way a workflow does, and these are the jobs teams bring most often. [Ask a question](https://app.sleuth.studio) The six workflows cover common decisions, but most of what teams bring is their own question. Ask it in Analyze in plain words. For anything bigger than a quick answer, Sleuth shows a plan and the columns it will build before it uses credits, just like a workflow. [Ask a question](ask-question.md) ## Common jobs, and how to ask | Job | Ask something like | | --- | --- | | Map a target, a disease, or a modality | "Map every program against [target] from preclinical to approved, with sponsor, stage, modality, and latest data." | | Track competitors and catalysts | "List the readouts expected in [indication] over the next 18 months, with the trial, the sponsor, and the date source." | | Benchmark clinical data | "Compare response rates and safety for every [class] drug in [indication], cohort by cohort, and flag comparisons that aren't like for like." | | Find deal comps | "Show licensing deals for [modality] assets at Phase 1 or 2 since 2021, with upfront, total value, and territory." | | Screen companies | "Find private biotechs with an unpartnered [modality] program in [area] and no financing in the last 12 months." | | Scan China | "List Chinese programs against [target] with their sponsor, stage, and whether rights outside China are available." [China coverage](china.md) | ## Write a question Sleuth can plan well - **Say the decision.** "So we can choose a lead indication" tells Sleuth what to weigh. - **Name the scope.** The target, indication, stages, regions, and years you care about. - **List the columns you'll want.** Sleuth suggests them, but naming yours gets you there in one pass. - **Add what you know.** A deck, a list of companies, or a spreadsheet to start from. [Bring your own documents and data](own-data.md) ## Keep it going Add a column to every row with a custom column, set the project to check for changes with monitoring, or check values cell by cell in Review mode. [Custom columns](custom-columns.md), [Monitoring](monitoring.md), [Review mode](review-mode.md) ## Related - [Ask a question in Analyze](ask-question.md): How Sleuth turns a question in your own words into a plan, datasets, and an answer you can check. - [Add your own columns](custom-columns.md): Ask Sleuth to fill in a column you define, on any dataset in Explore or Analyze. - [Prompts to start with](ai-prompts.md): Fill in the brackets and paste a prompt into your AI tool. Sleuth shows you its plan before it builds anything. *Updated November 9, 2026.* Source: md/wf-own-question.md --- # Sleuth MCP server Connect any MCP client or agent to Sleuth: run research, look up the knowledge base, build Explore datasets, and search the research library, with 52 tools. The Sleuth MCP server lets Claude, ChatGPT, Copilot, and any agent that speaks the Model Context Protocol use Sleuth. Every project an agent starts is saved to the user's Sleuth account, so they can open it in Studio, review it cell by cell, and keep working. ## Connection details | | | | --- | --- | | Server URL | `https://mcp.sleuthapi.com/mcp` | | Transport | Streamable HTTP | | Sign-in for people | OAuth 2.1 with PKCE. Clients register with dynamic client registration or a client ID metadata document. No API key needed. | | Sign-in for headless agents | A Sleuth API key sent as a bearer token. API keys come with Enterprise. [API](api.md) | | Plans | Every plan, including the free trial, for people who sign in. Jobs that run without a person need an API key, which comes with Enterprise. | | Directory listings | Coming soon for Claude and ChatGPT. Until then, add Sleuth as a custom connector with the server URL. | ## Add it to a client Claude Code: ``` claude mcp add --transport http sleuth https://mcp.sleuthapi.com/mcp ``` Clients that read an `mcpServers` config, like Cursor: ``` { "mcpServers": { "sleuth": { "url": "https://mcp.sleuthapi.com/mcp" } } } ``` The first call opens Sleuth's sign-in in a browser. Step-by-step guides: [Claude](setup-claude.md), [ChatGPT](setup-chatgpt.md), [Copilot](setup-copilot.md), [other tools](setup-other.md). ## The tools | Group | Tools | What they do | Credits | | --- | --- | --- | --- | | [Projects and runs](tools-projects.md) | 9 | Create a project, ask a research question, follow the run, approve plans, read the answer | Research runs use credits | | [Datasets and charts](tools-datasets.md) | 7 | Read datasets page by page, download CSV, list versions, get charts as specs or SVG | No | | [Your documents](tools-documents.md) | 5 | Upload files to a project, search them, get download links, delete them | No | | [Knowledge base lookups](tools-knowledge-base.md) | 15 | Resolve drugs, companies, indications, and trials to Sleuth IDs and follow their links | No | | [Explore datasets](tools-explore.md) | 4 | Build a filtered dataset from the knowledge base and download it | No | | [Research library](tools-research.md) | 4 | Search news, literature, filings, patents, and sell-side research, and read passages | No | | [Onboarding and health](tools-onboarding.md) | 8 | Set up a new account from the AI tool, and check the server | No | ## What it costs The same as Studio. A research run uses one credit per analytical cell it fills, and the user approves the plan first unless they turned on auto-approve. Lookups, Explore datasets, research library searches, and questions about results Sleuth already built don't use credits. [How credits work](how-credits-work.md) ## What isn't available over MCP yet - Workflows run in Studio. Agents can ask for the same analysis as a research question. [Prompts to start with](ai-prompts.md) - Monitoring, sharing, and plan changes are set up in Studio. ## Next [How a research run works](mcp-run-lifecycle.md) walks through a run from question to dataset. [Errors and limits](mcp-errors.md) lists every limit and the fix for each error. Writing an agent that reads these docs? See [Docs for AI agents](for-agents.md). ## Related - [How a research run works](mcp-run-lifecycle.md): Create a project, send the question, wait for the status to change, handle approvals and clarifications, then read the datasets and the answer. - [Sleuth in Claude, ChatGPT, and Copilot](ai-tools-overview.md): Ask in the AI tool you already use, and Sleuth does the work behind the answer. - [Docs for AI agents](for-agents.md): Every page in these docs has a Markdown version, there’s an index of all of them, and the whole set comes as one file. Here’s what’s where, and how agents should use Sleuth. *Updated November 9, 2026.* Source: md/mcp-overview.md --- # How a research run works Create a project, send the question, wait for the status to change, handle approvals and clarifications, then read the datasets and the answer. A research run is Sleuth's analysis for one question, the same one Studio runs when you ask in Analyze. It moves through a set of statuses. Your agent's job is to follow the status and do what each one asks. ## The run, step by step 1. **Create a project.** `create_project` with a name returns a `project_id` and status `idle`. Reuse the project for follow-up questions on the same topic. 2. **Attach files, if any.** `upload_document` before you send the question. Files over 10 MB go in through Studio. 3. **Send the question.** `send_message` with `message`. It returns a 202 acknowledgement at once, and the run starts. 4. **Wait for the status to change.** `wait_for_update` polls for you, for 180 seconds by default and up to 300 a call, and returns the full project state with `status_changed` and `waited_seconds`. Runs take longer as the plan grows, so keep chaining calls until the status changes. If you poll `get_project` yourself, respect its `retry_after` hint. 5. **Act on the status**, using the table below. 6. **Read the results.** `get_messages` for Sleuth's written answer, `list_datasets` and `get_dataset` for the data, and `list_artifacts` for charts. ## Statuses | Status | What's happening | What your agent does | | --- | --- | --- | | `idle` | The project exists and nothing is running. | Send a question. | | `processing` | Sleuth is reading the question and planning. | Wait. | | `awaiting_clarification` | Sleuth needs an answer before it can plan. | Read the latest message with `get_messages`, ask the user, and reply with `send_message`. | | `awaiting_plan_approval` | The plan is ready: the steps and the size of the job. | Show the user the plan. Send `approve_plan`, or `reject_plan` with `feedback`. | | `awaiting_schema_approval` | The columns are ready for review. | Show the user the columns. Send `approve_schema`, or `reject_schema` with `feedback`. | | `executing` | Sleuth is building the datasets. Credits are used here. | Wait. | | `complete` | Results are ready. | Read the answer, datasets, and charts. | | `failed` | The run stopped. A run that fails without results doesn't use credits. | Read the last message and try again, or tell the user to contact Sleuth. | | `cancelled` | Someone cancelled the run. | Nothing, unless the user wants to start again. | ## Approvals Whether a run pauses for approval depends on how the agent signed in: - **People signed in with OAuth, and personal API keys,** follow the user's auto-approve setting in Studio. With it off, runs pause at `awaiting_plan_approval` and `awaiting_schema_approval`. - **Organization API keys** approve both gates automatically, so runs go from `processing` to `executing` to `complete`. When a person is in the loop, show them the plan or the columns before you approve. A headless job on an organization key never pauses, so decide its scope before you send the question. There's no per-run credit cap yet: put the limits in the question itself, like "up to 50 rows and 3 analytical columns," and watch the Usage page. Approve with `send_message` and an `action`, never with a chat message: ``` send_message({ "project_id": "…", "action": "approve_plan" }) send_message({ "project_id": "…", "action": "reject_schema", "feedback": "Add a column for the latest readout date" }) ``` ## Cancel a run `send_message` with `"action": "cancel"` stops a run that hasn't finished. Cells already filled keep their credits, and nothing else is charged. ## Follow-up questions Send another `send_message` to the same project. Questions about results already built don't use credits. A follow-up that needs new rows or columns runs a new plan. ## Read results without flooding your context - `get_dataset` returns 50 rows by default and up to 2,000 a page. Use `row_offset` to page through, and `total_rows` to know when to stop. - `download_dataset_csv` returns the whole dataset as CSV text, so save it for when the user wants the file. - `export_project` returns everything in the project, with datasets capped at 10,000 rows each. Use it for archiving, not for reading. ## Related - [Projects and runs tools](tools-projects.md): Nine tools to create projects, ask research questions, follow runs, approve plans, and read Sleuth’s answers. - [Datasets and charts tools](tools-datasets.md): Seven tools to read the datasets and charts a run builds: page through rows, download CSV, compare versions, and get charts as specs or SVG. - [Plans and approvals](plans-approvals.md): Sleuth shows you what it will build before it builds it. Here is what you approve and how to skip approvals. *Updated November 9, 2026.* Source: md/mcp-run-lifecycle.md --- # Errors and limits Every limit an agent can hit on the Sleuth MCP server, and what to do about each error. ## Limits | What | Limit | | --- | --- | | File upload through MCP | 10 MB per file, decoded. Upload larger files in Studio. | | Rows per `get_dataset` call | 50 by default, up to 2,000 | | Rows per dataset in `export_project` | 10,000 | | One `wait_for_update` call | Up to 300 seconds. Chain calls for longer runs. | | Messages per `get_messages` call | Up to 200 | | Projects per `list_projects` call | Up to 100 | | Results per research library search | Up to 100 | | Passages per `query_research_document` call | Up to 50 | | Matches per `search_documents` call | Up to 50 | | Download links | Short-lived. Ask for a new one. | | Research runs at once, per API key | 10. More returns `run_limit_exceeded`. | | Explore datasets | Row caps by anchor. `explore_get_filters` lists them. | ## Errors and fixes | You see | What it means | Fix | | --- | --- | --- | | `onboarding_required` | The account isn't set up yet. | Call `get_onboarding_status` and follow `next_action`. [Onboarding](tools-onboarding.md) | | Waiting for approval | A new access request is with Sleuth. | Tell the user they'll get an email, then reconnect. | | Out of credits, trial ended, or payment failed | The workspace can't start new research. Sleuth's message says which, and links to plans or billing. | Show Sleuth's message as it is. [If you run out of credits](out-of-credits.md) | | `run_limit_exceeded` | An API key already has 10 research runs going. | Wait for one to finish, then send again. | | `409` or "project busy" | A research run is already going on that project. | Don't send again. Wait with `wait_for_update` until the run finishes, or send `cancel` if the user wants to stop it. | | "Can't reach the storage host" when fetching a link | The AI tool's sandbox can't reach the download address. That's the client, not Sleuth. | Give the user the link to open in a browser. | | `403` with an API key | API access isn't on for this organization. | Contact Sleuth. [API](api.md) | | `422` from `get_artifact_svg` | The artifact isn't a chart. | Use `get_artifact` instead. | | No matches right after an upload | The document is still being indexed. | Wait a minute and search again. | | Ambiguous filter value | A term matched more than one Explore value. | Ask the user which one they meant. | | Expired or broken download link | Links last minutes. | Call the tool again for a fresh link. | | Sell-side download refused | The organization doesn't have download access. | Read passages with `query_research_document`, or talk to Sleuth about access. | | Run status `failed` | The run stopped. It doesn't use credits. | Read the last message, then try again or contact Sleuth. | | Sign-in loops or "not allowed" | The client didn't finish OAuth, or the connection is stale. | Remove and re-add the connector, then sign in again. [Fix connection problems](ai-troubleshooting.md) | For anything else, the user can email support@sleuthinsights.com. ## Related - [Fix connection problems](ai-troubleshooting.md): The most common problems connecting Sleuth to an AI tool, and how to fix each one. - [How a research run works](mcp-run-lifecycle.md): Create a project, send the question, wait for the status to change, handle approvals and clarifications, then read the datasets and the answer. - [Get help](get-help.md): Where to ask, depending on what you need. *Updated November 9, 2026.* Source: md/mcp-errors.md --- # All 52 tools Every tool on the Sleuth MCP server in one table, with its group and whether it uses credits. Only one call uses credits: a research question sent with `send_message`, at one credit per analytical cell the run fills. A follow-up that needs new rows or analytical columns runs a new plan and uses credits too; questions about results already built don't. Everything else is free. | Tool | What it does | Group | Credits | | --- | --- | --- | --- | | `create_project` | Create a project | [Projects and runs](tools-projects.md) | No | | `send_message` | Ask a research question, answer a clarification, approve, reject, or cancel | [Projects and runs](tools-projects.md) | Research questions: 1 per analytical cell | | `get_project` | Get the project's state and status | [Projects and runs](tools-projects.md) | No | | `wait_for_update` | Wait until the status changes, up to 300 seconds a call | [Projects and runs](tools-projects.md) | No | | `get_messages` | Read the chat, including Sleuth's answers | [Projects and runs](tools-projects.md) | No | | `list_projects` | List projects | [Projects and runs](tools-projects.md) | No | | `update_project` | Rename a project | [Projects and runs](tools-projects.md) | No | | `delete_project` | Delete a project and everything in it | [Projects and runs](tools-projects.md) | No | | `export_project` | Export a whole project | [Projects and runs](tools-projects.md) | No | | `list_datasets` | List a project's datasets | [Datasets and charts](tools-datasets.md) | No | | `get_dataset` | Read rows a page at a time, up to 2,000 | [Datasets and charts](tools-datasets.md) | No | | `list_dataset_versions` | List a dataset's versions | [Datasets and charts](tools-datasets.md) | No | | `download_dataset_csv` | Get a whole dataset as CSV text | [Datasets and charts](tools-datasets.md) | No | | `list_artifacts` | List a project's charts | [Datasets and charts](tools-datasets.md) | No | | `get_artifact` | Get a chart's spec | [Datasets and charts](tools-datasets.md) | No | | `get_artifact_svg` | Get a chart as SVG | [Datasets and charts](tools-datasets.md) | No | | `upload_document` | Upload a file to a project, up to 10 MB | [Your documents](tools-documents.md) | No | | `list_documents` | List a project's documents | [Your documents](tools-documents.md) | No | | `get_document` | Get a document's details and a short-lived download link | [Your documents](tools-documents.md) | No | | `search_documents` | Search the text of your uploads | [Your documents](tools-documents.md) | No | | `delete_document` | Delete an uploaded document | [Your documents](tools-documents.md) | No | | `search_drugs` | Find drugs by name or code | [Knowledge base lookups](tools-knowledge-base.md) | No | | `get_drug_details` | Get a drug's details | [Knowledge base lookups](tools-knowledge-base.md) | No | | `find_drug_related_items` | Find a drug's targets, trials, indications, and more | [Knowledge base lookups](tools-knowledge-base.md) | No | | `find_drug_deals` | Find a drug's deals | [Knowledge base lookups](tools-knowledge-base.md) | No | | `find_drugs_by_target` | Find drugs by target | [Knowledge base lookups](tools-knowledge-base.md) | No | | `find_drugs_by_mechanism_of_action` | Find drugs by mechanism | [Knowledge base lookups](tools-knowledge-base.md) | No | | `find_drugs_by_modality` | Find drugs by modality | [Knowledge base lookups](tools-knowledge-base.md) | No | | `find_drugs_by_therapeutic_area` | Find drugs by therapeutic area | [Knowledge base lookups](tools-knowledge-base.md) | No | | `search_indications` | Find indications by name | [Knowledge base lookups](tools-knowledge-base.md) | No | | `find_indication_related_items` | Find drugs for indications | [Knowledge base lookups](tools-knowledge-base.md) | No | | `search_companies` | Find companies by name | [Knowledge base lookups](tools-knowledge-base.md) | No | | `find_company_related_assets` | Find a company's assets | [Knowledge base lookups](tools-knowledge-base.md) | No | | `search_clinical_trials` | Search trials | [Knowledge base lookups](tools-knowledge-base.md) | No | | `get_clinical_trial_by_code` | Get a trial by registry code | [Knowledge base lookups](tools-knowledge-base.md) | No | | `get_clinical_trial_attributes` | Get trial details | [Knowledge base lookups](tools-knowledge-base.md) | No | | `explore_get_filters` | List anchors, columns, filters, and row limits | [Explore datasets](tools-explore.md) | No | | `explore_search_filter_values` | Turn a user's term into a filter value | [Explore datasets](tools-explore.md) | No | | `explore_build_dataset` | Build and save a filtered dataset | [Explore datasets](tools-explore.md) | No | | `explore_download_dataset` | Download it as CSV or JSON | [Explore datasets](tools-explore.md) | No | | `search_research_documents` | Search news, literature, filings, patents, and sell-side research | [Research library](tools-research.md) | No | | `list_recent_research_documents` | List documents most recently added to a category | [Research library](tools-research.md) | No | | `query_research_document` | Pull passages from one sell-side or market intelligence document | [Research library](tools-research.md) | No | | `download_research_document` | Get a download link | [Research library](tools-research.md) | No | | `get_onboarding_status` | Get a new user's onboarding step | [Onboarding and health](tools-onboarding.md) | No | | `save_onboarding_role` | Save the user's role | [Onboarding and health](tools-onboarding.md) | No | | `save_onboarding_usage_mode` | Save where they'll use Sleuth | [Onboarding and health](tools-onboarding.md) | No | | `save_onboarding_use_cases` | Save the jobs they want done | [Onboarding and health](tools-onboarding.md) | No | | `save_onboarding_work_style` | Save how they work | [Onboarding and health](tools-onboarding.md) | No | | `submit_onboarding_access_request` | Request access from the chat | [Onboarding and health](tools-onboarding.md) | No | | `complete_mcp_setup` | Finish setup | [Onboarding and health](tools-onboarding.md) | No | | `get_health` | Check that Sleuth is up | [Onboarding and health](tools-onboarding.md) | No | Each group's page has every tool's inputs, defaults, and limits. ## Related - [Sleuth MCP server](mcp-overview.md): Connect any MCP client or agent to Sleuth: run research, look up the knowledge base, build Explore datasets, and search the research library, with 52 tools. - [How a research run works](mcp-run-lifecycle.md): Create a project, send the question, wait for the status to change, handle approvals and clarifications, then read the datasets and the answer. - [Errors and limits](mcp-errors.md): Every limit an agent can hit on the Sleuth MCP server, and what to do about each error. *Updated November 9, 2026.* Source: md/tools-all.md --- # Projects and runs tools Nine tools to create projects, ask research questions, follow runs, approve plans, and read Sleuth’s answers. Research runs are the only MCP calls that use credits. See [How a research run works](mcp-run-lifecycle.md) for the order to call these in. ### `create_project` Create a new research project. Returns `project_id` and status `idle`. | Input | Type | Notes | | --- | --- | --- | | `name` (required) | string | 1 to 255 characters | ### `send_message` Send a research question, an answer to a clarification, or an action to a project. Send exactly one of `message` or `action`. Returns a 202 acknowledgement right away. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | | `message` | string | The question or clarification answer | | `action` | string | `cancel`, `approve_plan`, `approve_schema`, `reject_plan`, or `reject_schema` | | `feedback` | string | Only with `reject_plan` or `reject_schema`. Tells Sleuth what to change. | To attach files, call `upload_document` first. Credits: a research question uses one credit per analytical cell the run fills. ### `get_project` Get the full project state, including its status. This is the main polling call. When the status is `awaiting_clarification`, read the latest message with `get_messages` and answer with `send_message`. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | Respect the `retry_after` hint, in seconds, between polls. ### `wait_for_update` Wait until the project’s status changes, or until the time budget runs out. Returns the latest state, plus `status_changed` and `waited_seconds`. Chain calls for long runs. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | | `timeout_seconds` | number | Default 180, maximum 300 | ### `get_messages` List the user and Sleuth messages in a project, including Sleuth’s written answers and questions. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | | `limit` | integer | 1 to 200, default 50 | | `offset` | integer | Default 0 | ### `list_projects` List the projects the signed-in user or key can see. | Input | Type | Notes | | --- | --- | --- | | `limit` | integer | 1 to 100, default 20 | | `offset` | integer | Default 0 | ### `update_project` Rename a project. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | | `name` (required) | string | The new name | ### `delete_project` Delete a project and everything in it. This can’t be undone, so confirm with the user first. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | ### `export_project` Export a full snapshot of a project. It can be very large, and datasets are capped at 10,000 rows each. Prefer the dataset and chart tools when an agent only needs part of it. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | ## Related - [How a research run works](mcp-run-lifecycle.md): Create a project, send the question, wait for the status to change, handle approvals and clarifications, then read the datasets and the answer. - [Datasets and charts tools](tools-datasets.md): Seven tools to read the datasets and charts a run builds: page through rows, download CSV, compare versions, and get charts as specs or SVG. - [Your documents tools](tools-documents.md): Five tools to add your own files to a project, search them, get download links, and remove them. *Updated November 9, 2026.* Source: md/tools-projects.md --- # Datasets and charts tools Seven tools to read the datasets and charts a run builds: page through rows, download CSV, compare versions, and get charts as specs or SVG. None of these use credits. ### `list_datasets` List the datasets in a project. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | ### `get_dataset` Get a dataset with its rows a page at a time. Use `total_rows`, `row_offset`, and `row_limit` to page through. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | | `dataset_id` (required) | string | Dataset UUID | | `row_limit` | integer | 1 to 2,000, default 50 | | `row_offset` | integer | Default 0 | | `version` | integer | A past version number. Omit for the latest. | | `include_csv` | boolean | Default false. True adds the same rows as CSV text. | ### `list_dataset_versions` List a dataset’s versions. Sleuth saves a new version when a dataset changes, for example after monitoring finds an update. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | | `dataset_id` (required) | string | Dataset UUID | ### `download_dataset_csv` Download the whole dataset as CSV text. Large datasets make large results, so use `get_dataset` when a sample will do. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | | `dataset_id` (required) | string | Dataset UUID | | `version` | integer | A past version number. Omit for the latest. | ### `list_artifacts` List the charts and other artifacts in a project. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | ### `get_artifact` Get an artifact in full. Its `artifact` field is the chart spec. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | | `artifact_id` (required) | string | Artifact UUID | ### `get_artifact_svg` Get an artifact rendered as SVG. Returns a 422 error for artifacts that aren’t visual. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | | `artifact_id` (required) | string | Artifact UUID | ## Related - [Projects and runs tools](tools-projects.md): Nine tools to create projects, ask research questions, follow runs, approve plans, and read Sleuth’s answers. - [Outputs and exports](outputs.md): What Sleuth hands back, and how to get each one out of Sleuth. - [Errors and limits](mcp-errors.md): Every limit an agent can hit on the Sleuth MCP server, and what to do about each error. *Updated November 9, 2026.* Source: md/tools-datasets.md --- # Your documents tools Five tools to add your own files to a project, search them, get download links, and remove them. Files you upload stay in the project and in your Sleuth workspace, and Sleuth reads them alongside the knowledge base when it runs. None of these tools use credits. [Bring your own documents and data](own-data.md) ### `upload_document` Upload a file to a project. Send the file’s bytes as base64. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | | `filename` (required) | string | The original file name, with its extension | | `content_base64` (required) | string | The file, base64-encoded. Up to 10 MB once decoded. | | `title` | string | A title for the document | Upload files before the `send_message` that should use them. Larger files go in through Studio. ### `list_documents` List the documents in a project. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | ### `get_document` Get a document’s details, including a `download_url` that expires within minutes. Fetch it right away, or present it to the user as a link. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | | `document_id` (required) | string | Document UUID | ### `search_documents` Search the text of documents uploaded to a project. A document uploaded moments ago may not match yet while Sleuth indexes it. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | | `query` (required) | string | A question or topic in plain words | | `document_ids` | string[] | Limit the search to these documents | | `top_k` | integer | 1 to 50, default 10 | ### `delete_document` Delete a document from a project. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | Project UUID | | `document_id` (required) | string | Document UUID | ## Related - [Use your own documents and data](own-data.md): Upload files for Sleuth to read, start from your own spreadsheet, or have Sleuth source the values you already have. - [Projects and runs tools](tools-projects.md): Nine tools to create projects, ask research questions, follow runs, approve plans, and read Sleuth’s answers. - [Your data in AI tools](ai-tools-data.md): What travels between your AI tool and Sleuth when you connect them. *Updated November 9, 2026.* Source: md/tools-documents.md --- # Knowledge base lookup tools Fifteen tools to resolve drugs, companies, indications, and trials to Sleuth IDs, read their details, and follow their links to deals, targets, and programs. These read Sleuth's knowledge base directly and don't use credits. Most follow the same pattern: search by name to get an ID, then pass the ID to a details or related-items tool. Never guess an ID. Results come back a page at a time: pass `offset` to get the next page. ## Drugs ### `search_drugs` Find drugs or assets by primary name or synonym, and get their Sleuth IDs. | Input | Type | Notes | | --- | --- | --- | | `drug_name` (required) | string | A name, code, or synonym, such as "enzomenib" or "DSP-5336" | ### `get_drug_details` Get structured details for drug IDs returned by `search_drugs`. | Input | Type | Notes | | --- | --- | --- | | `drug_ids` (required) | string[] | One or more drug IDs | ### `find_drug_related_items` Find the mechanisms, modalities, therapeutic areas, trials, indications, or deals linked to drugs. | Input | Type | Notes | | --- | --- | --- | | `drug_ids` (required) | string[] | One or more drug IDs | | `relation_type` | string[] | Limit to some kinds of links | | `offset` | integer | Default 0 | ### `find_drug_deals` Find deals involving drugs. | Input | Type | Notes | | --- | --- | --- | | `drug_ids` (required) | string[] | One or more drug IDs | | `offset` | integer | Default 0 | ### `find_drugs_by_target` Find drugs that act on a target, such as PD-1 or GLP-1R. | Input | Type | Notes | | --- | --- | --- | | `target` (required) | string | The target name | | `offset` | integer | Default 0 | ### `find_drugs_by_mechanism_of_action` Find drugs by mechanism of action, such as GLP-1R agonist. | Input | Type | Notes | | --- | --- | --- | | `moa` (required) | string | The mechanism | | `offset` | integer | Default 0 | ### `find_drugs_by_modality` Find drugs with a modality, such as monoclonal antibodies. | Input | Type | Notes | | --- | --- | --- | | `modality` (required) | string | The modality | | `offset` | integer | Default 0 | ### `find_drugs_by_therapeutic_area` Find drugs in a therapeutic area, such as diabetes or cancer. | Input | Type | Notes | | --- | --- | --- | | `therapeutic_area` (required) | string | The therapeutic area | | `offset` | integer | Default 0 | ## Indications ### `search_indications` Find indications by name and get their Sleuth IDs. | Input | Type | Notes | | --- | --- | --- | | `indication_name` (required) | string | The indication, such as "NMOSD" | ### `find_indication_related_items` Find drugs linked to indications. | Input | Type | Notes | | --- | --- | --- | | `indication_ids` (required) | string[] | One or more indication IDs | | `offset` | integer | Default 0 | ## Companies ### `search_companies` Find companies by name and get their Sleuth IDs. | Input | Type | Notes | | --- | --- | --- | | `company_name` (required) | string | The company name | ### `find_company_related_assets` Find a company’s assets, optionally by the kind of link (developer, licensee, and so on). | Input | Type | Notes | | --- | --- | --- | | `company_ids` (required) | string[] | One or more company IDs | | `relation_type` | string[] | Limit to some kinds of links | | `offset` | integer | Default 0 | ## Clinical trials ### `search_clinical_trials` Search clinical trials. Criteria are combined with AND, and a value can list alternatives with OR. | Input | Type | Notes | | --- | --- | --- | | `queries` (required) | object[] | Each item has `attribute` (`title`, `keywords`, `acronym`, `condition`, or `location`) and `value` | ### `get_clinical_trial_by_code` Get a trial by its registry code, such as NCT01234567. | Input | Type | Notes | | --- | --- | --- | | `code` (required) | string | The registry code | ### `get_clinical_trial_attributes` Get structured attributes for trial IDs. | Input | Type | Notes | | --- | --- | --- | | `clinical_trial_ids` (required) | string[] | One or more trial IDs | ## Related - [What's in the knowledge base](knowledge-base.md): The records Sleuth works from, how many there are, and how they connect. - [Explore tools](tools-explore.md): Four tools to build a filtered dataset from the knowledge base, the same way Explore does in Studio, and download it. - [Every source](sources.md): The registries, regulators, journals, filings, and feeds Sleuth reads, grouped by what they answer. *Updated November 9, 2026.* Source: md/tools-knowledge-base.md --- # Explore tools Four tools to build a filtered dataset from the knowledge base, the same way Explore does in Studio, and download it. Explore datasets come straight from the knowledge base, so they don't use credits. The dataset is saved to a project in Studio, where the user can open it, add columns, or set up monitoring. [Explore the knowledge base](explore.md) ## The order to call them in 1. `explore_get_filters` to see the anchors, columns, filters, allowed values, and row limits. 2. `explore_search_filter_values` for every term the user gave, like PD-1, NSCLC, or ADC. Pass the values it returns unchanged. If a term matches more than one value, ask the user which one. 3. `explore_build_dataset` with the resolved filters. Show the user the counts and preview. 4. Ask whether they want the full dataset. If they do, `explore_download_dataset`, as CSV unless they asked for JSON. Use this tool, not `download_dataset_csv`, for Explore datasets: its CSV matches a Studio export column for column. ### `explore_get_filters` List the dataset anchors (the kind of record each row is), the columns you can pick, the filters, their fixed allowed values, and the row limits. Call it before building a dataset. *No inputs.* ### `explore_search_filter_values` Turn a user’s term into the canonical filter values Explore uses. Never guess IDs. | Input | Type | Notes | | --- | --- | --- | | `category` (required) | string | The filter category, from `explore_get_filters` | | `query` (required) | string | The user’s term, such as "PD-1" | | `limit` | integer | Default 20 | ### `explore_build_dataset` Build and save a dataset from resolved filter values. Returns counts, columns, a short preview, download handles, any row-cap warnings, and a link to the dataset in Studio. | Input | Type | Notes | | --- | --- | --- | | `anchor` (required) | string | The kind of record each row is, from `explore_get_filters` | | `filters` (required) | object | Filter category to a list of values, each with the `label` and the `entity_guid` or `entity_guids` returned by `explore_search_filter_values` | | `columns` | string[] | The columns to include. Omit for the defaults. | | `title` | string | A name for the dataset | | `preview_rows` | integer | Default 5 | | `generate_anyway` | boolean | Default false. Set to true to build the dataset after a row-cap warning. | ### `explore_download_dataset` Download the full saved dataset. CSV returns a short-lived download link with the same columns as a Studio export. JSON returns every row, which makes a large result for big datasets. | Input | Type | Notes | | --- | --- | --- | | `project_id` (required) | string | From `explore_build_dataset` | | `dataset_id` (required) | string | From `explore_build_dataset` | | `format` | string | `csv` (default) or `json` | ## Related - [Explore the knowledge base](explore.md): Filter the records Sleuth works from and build your own dataset, with no credits. - [Knowledge base lookup tools](tools-knowledge-base.md): Fifteen tools to resolve drugs, companies, indications, and trials to Sleuth IDs, read their details, and follow their links to deals, targets, and programs. - [Datasets and charts tools](tools-datasets.md): Seven tools to read the datasets and charts a run builds: page through rows, download CSV, compare versions, and get charts as specs or SVG. *Updated November 9, 2026.* Source: md/tools-explore.md --- # Research library tools Four tools to search the documents Sleuth collects, from news to sell-side research, list the newest in a category, read the passages that matter, and get download links. These search the same library as document search in Studio and don't use credits. [Search documents](document-search.md) ## Categories | Category ID | What's in it | | --- | --- | | `news` | News and press releases | | `scientific-literature` | Journal articles, preprints, and conference abstracts | | `regulatory` | Regulatory documents | | `company-filings` | Company filings | | `patents` | Patents | | `broker-research` | Sell-side research | | `company-presentations` | Company presentations and transcripts | | `other-documents` | Other documents | Searches cover the categories your account can see. Downloading sell-side research and market intelligence needs download access on your organization. Without it, you can still read the relevant passages with `query_research_document`. [Sell-side research](sell-side.md) ### `search_research_documents` Search the research library. Omit `categories` to search every category your account can see. | Input | Type | Notes | | --- | --- | --- | | `query` (required) | string | A question or topic in plain words | | `categories` | string[] | Category IDs from the table above | | `top_k` | integer | 1 to 100, default 50 | ### `list_recent_research_documents` List the newest documents in one category, optionally from the last day, week, or month. | Input | Type | Notes | | --- | --- | --- | | `category` (required) | string | A category ID | | `limit` | integer | 1 to 100, default 50 | | `time_window` | string | `day`, `week`, or `month`. Omit for all history. | ### `query_research_document` Pull the passages that answer a question from one sell-side research or market intelligence document. | Input | Type | Notes | | --- | --- | --- | | `category` (required) | string | The category returned by search or the recent list | | `document_id` (required) | string | From search or the recent list | | `query` (required) | string | What to find in the document | | `top_k` | integer | 1 to 50, default 20 | ### `download_research_document` Get a download link for a document. It returns a link only, so present it as a link rather than saying the file was downloaded. | Input | Type | Notes | | --- | --- | --- | | `category` (required) | string | The category returned by search or the recent list | | `document_id` (required) | string | From search or the recent list | | `source_url` | string | The `source_url` from a search result, for documents held outside Sleuth | ## Related - [Search documents](document-search.md): Search Sleuth's document library, read any result in full, and see exactly where a value came from. - [Sell-side research](sell-side.md): Research from 35+ banks and brokers, included in every plan, searchable next to everything else. - [Knowledge base lookup tools](tools-knowledge-base.md): Fifteen tools to resolve drugs, companies, indications, and trials to Sleuth IDs, read their details, and follow their links to deals, targets, and programs. *Updated November 9, 2026.* Source: md/tools-research.md --- # Onboarding and health tools Eight tools that set up a new account from inside an AI tool, step by step, and check that the server is up. Someone who connects Sleuth to an AI tool before they have an account can set up their account without leaving the chat. Before a new user's first task, or when any tool returns `onboarding_required`, call `get_onboarding_status` and follow only the `current_step` and `next_action` it returns. People who already use Studio skip most of this. ## The steps | `current_step` | What the user does | Tool | | --- | --- | --- | | `role` | Confirms their job title, and optionally a LinkedIn URL and a description of their role | `save_onboarding_role` | | `setup` | Chooses where they'll use Sleuth: Studio, their AI tool, or both | `save_onboarding_usage_mode` | | `use_cases` | Picks and orders the jobs they want done | `save_onboarding_use_cases` | | `work_style` | Answers a few questions about how they work | `save_onboarding_work_style` | | `access_request` | Requests access, in Studio or right in the chat | `submit_onboarding_access_request` | | Browser steps | Accepts policies or joins a workspace in the browser, then reconnects the AI tool | None: ask the user to refresh the connection, then check status again | | `mcp_setup` | Confirms the connection works | `complete_mcp_setup` | When `next_action` is `await_approval`, the request is with Sleuth. Tell the user it needs approval and that they'll get an email. > **For agents:** Confirm every value with the user before you save it, and never fill in a value they didn't give. Use plain, customer-facing words, and don't tell the user what role or access their account has. ### `get_onboarding_status` Get the user’s onboarding state: `current_step`, `next_action`, and anything already filled in. Call it before a new user’s first task and after any `onboarding_required` error. *No inputs.* ### `save_onboarding_role` Save the user’s role. Use only when `current_step` is `role`. | Input | Type | Notes | | --- | --- | --- | | `job_title` (required) | string | Confirmed with the user | | `linkedin_url` | string | | | `role_description` | string | | ### `save_onboarding_usage_mode` Save where the user will use Sleuth. Use only when `current_step` is `setup`. | Input | Type | Notes | | --- | --- | --- | | `usage_mode` (required) | string | `studio`, `mcp`, or `both` | ### `save_onboarding_use_cases` Save the user’s jobs in priority order, using the IDs from `state.use_cases`. Use only when `current_step` is `use_cases`. | Input | Type | Notes | | --- | --- | --- | | `use_cases` (required) | string[] | IDs in the user’s order | ### `save_onboarding_work_style` Save the user’s work-style answers, using the question and option IDs from `get_onboarding_status`. Leave out questions they didn’t answer. Use only when `current_step` is `work_style`. | Input | Type | Notes | | --- | --- | --- | | `answers` (required) | object[] | Each with `question_id`, `selected` option IDs, and optional `other` text | ### `submit_onboarding_access_request` Submit an access request from the chat. Use only when `current_step` is `access_request` and the user chose to request it here instead of in Studio. | Input | Type | Notes | | --- | --- | --- | | `company_name` (required) | string | Confirmed with the user | | `country` (required) | string | Confirmed with the user | | `company_website` | string | Needed when the user signed in with a work email | | `linkedin_url` | string | Needed when the user signed in with a personal email | | `first_name` | string | Only if status didn’t return it | | `last_name` | string | Only if status didn’t return it | ### `complete_mcp_setup` Finish setup once the user confirms the connection works. Use only when `current_step` is `mcp_setup`. *No inputs.* ### `get_health` Check that the server and Sleuth’s API are up. It still needs a signed-in connection. *No inputs.* ## Related - [Sign up and get access](sign-up.md): How to request an account, what approval looks like, and what onboarding asks. - [Fix connection problems](ai-troubleshooting.md): The most common problems connecting Sleuth to an AI tool, and how to fix each one. - [Errors and limits](mcp-errors.md): Every limit an agent can hit on the Sleuth MCP server, and what to do about each error. *Updated November 9, 2026.* Source: md/tools-onboarding.md --- # API Sleuth’s REST API lets your own software run Sleuth research without the Studio screen. It comes with Enterprise and powers the MCP server. [Talk to us about API access](https://www.sleuthinsights.com/talk-to-us) Most teams don't need the API: agents and AI tools connect through the [MCP server](mcp-overview.md) on any plan. The API is for applications and internal systems that call Sleuth directly, like a pipeline that refreshes a dataset every night or a portal that shows Sleuth results. ## What it does Everything the MCP server's project tools do, as REST calls: - Create, list, rename, and delete projects, and send questions. - Follow a run's status, and approve, reject, or cancel at the plan and column steps. - Upload documents, including files over 10 MB, which the MCP server can't take. - Fetch datasets (also as CSV), charts (as SVG), and a whole project in one call. ## Keys | Key | Acts as | Approvals | | --- | --- | --- | | Organization key | Your organization | Plans and columns are approved automatically. | | Personal key | One Sleuth user, so runs show up as theirs in Studio | Follows that user's auto-approve setting. | Sleuth creates keys for you. Send the key as a bearer token. The same key also works on the MCP server, for agents that run without a person signing in. ## Getting access API access is turned on per organization, with Enterprise. Your Sleuth contact sends the keys and the API reference. [Enterprise](enterprise.md) ## Credits The same as Studio and the MCP server: one credit per analytical cell a run fills. [How credits work](how-credits-work.md) ## Related - [Enterprise](enterprise.md): What Enterprise adds for companies rolling Sleuth out across teams and systems. - [Sleuth MCP server](mcp-overview.md): Connect any MCP client or agent to Sleuth: run research, look up the knowledge base, build Explore datasets, and search the research library, with 52 tools. - [Security overview](security.md): The short answers for your security team, and where to find the evidence. *Updated November 9, 2026.* Source: md/api.md --- # Docs for AI agents Every page in these docs has a Markdown version, there’s an index of all of them, and the whole set comes as one file. Here’s what’s where, and how agents should use Sleuth. ## Machine-readable docs | What | Where | Use it for | | --- | --- | --- | | Index | `/llms.txt` | Every page, its Markdown link, and one line on what it answers. Start here. | | One page | Add `.md` to a page's URL | A single page as clean Markdown | | Everything | `/llms-full.txt` | All pages in one file, for search or retrieval | Every page also has **Copy page**, which copies it as Markdown to paste into a chat. ## Instructions for agents using Sleuth 1. **Use the MCP server to act, and these docs to explain.** Server: `https://mcp.sleuthapi.com/mcp`. [MCP server](mcp-overview.md) 2. **Check onboarding first for new users.** Call `get_onboarding_status` before a new user's first task, and follow its `next_action`. 3. **Show the plan before you approve it.** Research runs use the user's credits. When a person is in the loop, approve on their behalf only if they've told you to. A headless job on an organization key is approved automatically, so set its scope up front. 4. **Resolve every name to an ID.** Use the search tools, and never guess IDs. If a term is ambiguous, ask. 5. **Page through data.** Read datasets with `get_dataset` a page at a time instead of downloading whole exports into your context. 6. **Give links as links.** Download links expire within minutes. Don't say a file was downloaded unless you fetched it. 7. **Keep sources and confidence with values.** When you report a value from Sleuth, say where it came from and how confident Sleuth is. [Confidence scores](confidence.md) 8. **Send account problems to Sleuth.** For billing, access, or errors you can't fix, the user can email support@sleuthinsights.com. ## How these pages are written - One topic per page. The first paragraph answers the page's main question. - Exact names: tool names, statuses, and limits match the server. - Credit facts are on [How credits work](how-credits-work.md) and [Plans and pricing](plans.md). - Terms are defined in the [Glossary](glossary.md). ## Related - [Sleuth MCP server](mcp-overview.md): Connect any MCP client or agent to Sleuth: run research, look up the knowledge base, build Explore datasets, and search the research library, with 52 tools. - [How a research run works](mcp-run-lifecycle.md): Create a project, send the question, wait for the status to change, handle approvals and clarifications, then read the datasets and the answer. - [Glossary](glossary.md): The words you'll see in Sleuth and in these docs. *Updated November 9, 2026.* Source: md/for-agents.md --- # What’s new New features and changes in Sleuth, newest first. ## November 2026 ### Self-serve plans and a free trial *At launch* Start free with a 14-day trial: 2,500 credits and everything in Starter, with no card. Then choose Starter for one person or Team for a group, both paid in credits that roll over. [Free trial](free-trial.md), [Plans and pricing](plans.md) ### A help center for people and AI agents *At launch* These docs cover every part of Sleuth. Every page also comes as Markdown, with an index for AI tools at `/llms.txt`. [Docs for AI agents](for-agents.md) ### Sign up from your AI tool *November 7* Connect Sleuth in Claude or another AI tool without an account, and it walks you through setting one up in the chat. [Sign up and get access](sign-up.md) ### Monitor as many projects as you like *November 3* The limit on monitored projects is gone. Monitoring still uses credits only when it finds something new. [Monitoring](monitoring.md) ### Explore datasets from Claude and other AI tools *November 3* Describe the filters in plain words and your AI tool builds the Explore dataset: counts, a preview, and a download link, saved to a project in Studio. [Explore tools](tools-explore.md) ## October 2026 ### Sign up in minutes *October 29* Request access with your email, sign in with a one-time code, and onboarding sets up your workspace. [Sign up and get access](sign-up.md) ### Find drugs by disease, subtypes included *October 28* Explore's disease filter now follows a tree of medical concepts. Pick a disease and its subtypes come with it, and different names for the same disease return the same drugs. [Explore the knowledge base](explore.md) ### Verify a table you built yourself *October 23* Upload a table you made by hand and ask Sleuth to verify it. Sleuth finds sources for the values already in it without changing them, so you can see which ones hold up. [Bring your own documents and data](own-data.md) ### A new Analyze chat *October 19* Pins keep results in view, a Files tab holds your uploads and outputs, a task panel shows what's running, and suggested follow-ups are one click away. [Ask a question](ask-question.md) ### Document search in AI tools *October 15* Claude and other AI tools can search Sleuth's research library, list the documents most recently added to a category, and pull the passages that answer a question. [Research library tools](tools-research.md) ### The Usage page *October 13* Admins see what used the workspace's credits by event, project, and person, filter by date, and export the report. [The Usage page](usage-page.md) ### Monitoring settings, and dataset sharing on projects *October 9* More control over how monitoring works on each project, and everyone on a project can now share its datasets, not only admins. [Monitoring](monitoring.md), [Share your work](share-work.md) ### Studio on phones and tablets *October 7* Check a project or read results away from your desk, in your phone's browser. [Find your way around Studio](studio-tour.md) ### Custom columns *October 1* Ask a question of every row, and Sleuth fills in a new column with a value, its reasoning, a confidence score, and sources for each cell. [Add your own columns](custom-columns.md) ## September 2026 ### Your context *September 23* Sleuth remembers what you tell it about you and your company, from onboarding and your uploads, and every workflow and chat reads from it. You can see and change it at any time. [Your context](your-context.md) ## August 2026 ### Explore views *August 10* Start Explore from a ready-made view, then filter, pick columns, and save the dataset to a project. [Explore the knowledge base](explore.md) ## July 2026 ### Sign in from AI tools, and personal API keys *July 21* Connect Sleuth to your AI tool with a normal sign-in, no key needed. For Enterprise, personal API keys run work as a named user, so it shows up as theirs in Studio. [Sleuth in AI tools](ai-tools-overview.md), [API](api.md) ### Initiatives *July 20* Group related projects and their datasets in one folder, like everything for one asset or one deal. [Projects, initiatives, and sharing](share-work.md) ### Review mode, and a single-tenant option for Enterprise *July 15* Every value Sleuth fills has a confidence score and its sources, and Review mode lets you check, correct, and approve values cell by cell before you share. Enterprise can also run Sleuth in a dedicated environment of its own. [Review mode](review-mode.md), [Security overview](security.md) ### Sleuth API *July 7* A REST API for Enterprise customers' own software: create projects, ask questions, and fetch datasets and charts without the Studio screen. [API](api.md) ## Coming next **Extra usage.** A setting that lets a workspace keep working past its plan's credits, up to a limit an admin sets, is coming in December. [If you run out of credits](out-of-credits.md) **Sleuth in Claude's and ChatGPT's directories.** Listings you can add with one click are coming soon. Until then, add Sleuth as a custom connector in about a minute. [Set up Sleuth in Claude](setup-claude.md), [Set up Sleuth in ChatGPT](setup-chatgpt.md) ## Related - [Welcome to Sleuth](welcome.md): What Sleuth does, where you use it, and where to go in these docs for each job. - [Docs for AI agents](for-agents.md): Every page in these docs has a Markdown version, there’s an index of all of them, and the whole set comes as one file. Here’s what’s where, and how agents should use Sleuth. - [Plans and pricing](plans.md): Starter for one person, Team for any number of people, Enterprise for company-wide use, and what a credit costs. *Updated November 9, 2026.* Source: md/whats-new.md