> For the complete documentation index, see [llms.txt](https://help.dscout.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.dscout.com/dscout-ai/dscout-mcp.md).

# Dscout MCP

{% hint style="info" %}
**Note:** The Dscout MCP is currently in closed beta, meaning only certain users have access and the feature will undergo changes prior to general availability. For more information, contact your Dscout Account Director.
{% endhint %}

The Dscout MCP unlocks your study data and lets your favorite AI tools use it as real-time context, bringing your research projects inline with your AI workflows. Say goodbye to copying responses and pasting them into Claude for help with analysis. Gone are the days of building prototypes in one tool and designing research around them in another. With the Dscout MCP, you can do it all in the tools you're already using.

Some key use cases for the Dscout MCP include:

* **Study building and management:** Work with the Dscout MCP to design and build a Dscout study that meets your research goals. Define your recruitment criteria, provide any prototypes or concepts you want to test, and draft a launch-ready study right in your AI workflow.
* **Mission recaps and summaries:** Need a quick rundown of the results from a recent study you ran? With the Dscout MCP, you can get a high-level summary of key takeaways or action items based on real participant responses.
* **Artifact creation:** Want to turn your study results into shareable PDFs or visuals? The Dscout MCP can help by leveraging the asset-creation capabilities of your favorite AI tools.
* **Cross-mission analysis:** The Dscout MCP can help you compare responses from multiple Dscout studies at once. Simply tell the MCP which studies you want to analyze, then start asking questions.

### Dscout MCP vs. other Dscout AI tools

While part of Dscout AI, the Dscout MCP is a different way to work with your Dscout studies outside of the Dscout platform:

* **Other Dscout AI tools:** Tools like Explore your data and Draft your study live directly in the Dscout platform. You use these for a more Dscout-guided experience building and analyzing studies.
* **Dscout MCP:** The MCP connects external tools like Claude, Figma Make, and more directly to your Dscout data. A connection like this lets you use this data in other workflows alongside other sources, like Notion and Google Drive. For example, query Dscout from Figma Make, a tool where you're already building a prototype, so the feedback you gain from a Dscout study lives right where your prototype lives.

Use the Dscout MCP when you want to bring your data into multi-source AI workflows or support high-volume, automated analysis from external tools you're already working in.

Use other Dscout AI tools that are built into the Dscout platform when you want an interactive AI experience for more guided research and analysis.

### MCP approval steps

Many organizations require you to go through prior approval before you can use external MCP tools like Dscout's. Here are some tips to help you start the conversation:

* **Know who to talk to:** Required approval parties vary by organization but you'll most likely need to talk to your Security team, your IT/AI Admin team, and your Dscout Account Owner.
* **Know who needs what access:** When making your case, be clear about whether your whole team needs access to the MCP or if you want to start with a pilot group. Additionally, do you want to start with read-only access or do you need write-access too?
* **Have your resources handy:** Be prepared to share this document and the [Set up the Dscout MCP](/dscout-ai/dscout-mcp/set-up-the-dscout-mcp.md) document with the people listed above. Keep in mind that the person configuring the MCP will also need a Dscout seat to test things out, so be ready to help someone gain access to your Dscout account who may not normally have it.

The above tips should keep things moving within your organization, but Dscout is always here to help: If you run into any issues or if you have deeper questions, reach out to your Dscout Account Team for answers.

### Security overview

Here are some common questions you (or your IT team!) might have about how the Dscout MCP keeps your data secure:

<details>

<summary>What tools and permissions does the Dscout MCP have?</summary>

The Dscout MCP exposes a defined set of research-based tools to help users design and manage studies, query and summarize responses, and more. **The MCP is limited to permissions the user has within the Dscout platform, meaning no new capabilities or data are exposed.** For a complete list of tools built into the Dscout MCP, see [Dscout MCP tools](/dscout-ai/dscout-mcp/dscout-mcp-tools.md).

Read/write permissions are customizable, and you can choose which to allow—even if that means keeping the Dscout MCP read-only.

The MCP has some limitations that cannot be bypassed:

* Cannot build studies recruiting from Partner Panels or Private Panels.
* Cannot build AI moderated studies, media surveys, intercept studies, diary studies, interview studies, or standalone screeners. (However, the MCP does support analysis of these study types.)

</details>

<details>

<summary>What data is exposed by the Dscout MCP?</summary>

The only data the MCP can access is the data the authenticated user already has access to in Dscout. So, if a user doesn't have access to a specific project directly in Dscout, they won't be able to access it using the MCP either. Access is scoped to an individual user's account and role on every request.

</details>

<details>

<summary>How does the MCP authenticate and authorize access?</summary>

The Dscout MCP uses OAuth 2.0 with PKCE, browser-based sign-in, meaning users authenticate using their existing Dscout credentials. If your organization requires SSO or MFA, those requirements are upheld by the MCP. No API keys or secret keys are shared, stored, or needed.

Access tokens are short-lived and user-scoped: a **24-hour access token** plus a **30-day refresh token**. Only hashes are stored server-side; raw tokens are shown once, at the moment of authentication and not again.

**Entitlements are inherited from the user's Dscout role and enforced per request.** For example, a *viewer* can query a study but cannot launch one.

</details>

<details>

<summary>How is access managed and/or revoked?</summary>

Sessions are token-based and user-scoped, meaning they are tied to an individual and not an organization's overall Dscout account. Account-level authorization is rechecked on every request, so access always reflects the user's current Dscout permissions.

If a user's account is deactivated in the overall Dscout account, that user immediately loses access via the Dscout MCP as well.

</details>

<details>

<summary>Is the Dscout MCP security compliant?</summary>

Dscout is aligned with several security standards and undergoes annual audits to ensure we stay that way. For more information on specific certifications Dscout upholds, see [AI privacy and security in Dscout](/dscout-ai/what-is-dscout-ai/ai-privacy-and-security-in-dscout.md).

</details>

### Limitations

The Dscout MCP can do a lot but currently has a few limitations. The Dscout MCP cannot:

* Build studies recruiting from Partner Panels or Private Panels.
* Include rich-text formatting in question prompts, study overviews, etc.
* Build media surveys, AI moderated studies, intercept studies, diary studies, interview studies, or standalone screeners. (However, the MCP does support read-access for these study types.)

### Frequently asked questions (FAQ)

<details>

<summary>Can users with <code>Viewer</code> permissions use the MCP?</summary>

Yes! However, Viewers are still restricted to the same level of permissions they have in the standard Dscout app, meaning they can use the MCP in a read-only capacity and cannot use it to create studies.

</details>

### Learn more

Ready to get started? Check out the articles below for more information:

<table data-view="cards"><thead><tr><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td>Set up the Dscout MCP</td><td><a href="/pages/RPzs4FWA9RhDpS3vNJga">/pages/RPzs4FWA9RhDpS3vNJga</a></td></tr><tr><td>MCP prompt library</td><td><a href="/pages/fTkpJNxxmqP5IDa8tza5">/pages/fTkpJNxxmqP5IDa8tza5</a></td></tr><tr><td>Best practices</td><td><a href="/pages/4pCY26f1e7dGWINEtdKy">/pages/4pCY26f1e7dGWINEtdKy</a></td></tr><tr><td>MCP tool breakdown</td><td><a href="/pages/8CP7RJJXKxiUKXX3yeUK">/pages/8CP7RJJXKxiUKXX3yeUK</a></td></tr><tr><td>Troubleshooting</td><td><a href="/pages/1FCshUqTpPs6R4p4v8zD">/pages/1FCshUqTpPs6R4p4v8zD</a></td></tr></tbody></table>


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://help.dscout.com/dscout-ai/dscout-mcp.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.