> For the complete documentation index, see [llms.txt](https://docs.tonic.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.tonic.ai/fabricate/data-agent-tools/project-api/mcp-servers.md).

# Using MCP servers

You can create mock MCP servers that can call your mock API endpoints. You can then test an AI agent against the mock MCP server.

To interact with simulation results, Fabricate also provides a set of [built-in MCP servers for the simulation content platforms](/fabricate/databases-and-data/simulations/simulation-mcp-servers.md). You publish the built-in MCP servers that you want to use. For example, if your simulation results include Slack messages, then you can publish and use the Slack MCP server.

## Creating a mock MCP server <a href="#mcp-server-create" id="mcp-server-create"></a>

To create a mock MCP server, prompt the Data Agent to create the server. In the prompt, you specify the endpoints that you want to include.

For example, to create an MCP server that supports all of your mock API endpoints:

`Create a custom MCP server that can call all of the endpoints.`

The agent creates the MCP server and generates the corresponding tools for the supported requests.

## Viewing MCP server details <a href="#mcp-server-details" id="mcp-server-details"></a>

Under **API**, to display your MCP servers, click **MCPs**. The link displays the number of MCP servers.

<figure><img src="https://4109733485-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FmoU4gTR9LxlzHeWmQCUZ%2Fuploads%2FLtXgUIS7f9CZdlkDGekQ%2FDataAgent_MCPsLink.png?alt=media&amp;token=67bfd5e0-8d31-48c0-b966-d4d7361b2744" alt=""><figcaption><p>API menu with the MCPs link that displays the number of MCP servers</p></figcaption></figure>

For each MCP server, the **MCPs** page includes:

* A summary description of the MCP server.
* A list of URLS to connect to the MCP server. Includes:
  * Direct URLs for each version of the project.
  * Sandbox URLs for each version of the project.
* The list of tools in the MCP server.

<figure><img src="https://4109733485-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FmoU4gTR9LxlzHeWmQCUZ%2Fuploads%2FqEfdawN8qq8nyeXmRpD0%2FProject_MCPsPage.png?alt=media&amp;token=7a8d3a10-49d5-4ba1-99a3-acb3b1daf456" alt=""><figcaption><p>MCPs page with the mock MCP servers</p></figcaption></figure>

## Connecting an AI agent to an MCP server <a href="#mcp-server-connect-to" id="mcp-server-connect-to"></a>

To connect to an MCP server, you point the agent to an MCP server URL, either a project version or a project version sandbox.

The request must include a Fabricate API key as a bearer token in the Authorization header. Make sure that the API key has sufficient permission to perform the tasks that the MCP server supports.

To display the text to copy into your MCP client configuration, click **How to authenticate**.

## Deleting MCP servers and server tools <a href="#mcp-server-delete" id="mcp-server-delete"></a>

To delete a specific mock MCP server tool, on the **MCPs** page, click the **Delete** option for the tool. Note that for built-in mock MCP servers for simulation results, you cannot delete individual tools.

To delete an entire mock MCP server, click the **Delete** option for the MCP server.


---

# 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 by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.tonic.ai/fabricate/data-agent-tools/project-api/mcp-servers.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

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.