> ## Documentation Index
> Fetch the complete documentation index at: https://docs.beyondnoise.io/llms.txt
> Use this file to discover all available pages before exploring further.

# MCP Integration

> Connect AI clients like Claude Desktop and Cursor to BeyondNoise via the Model Context Protocol

BeyondNoise exposes an MCP (Model Context Protocol) server that lets AI clients like Claude Desktop, Cursor, and opencode interact with your workspaces, research, and brand context programmatically.

## How it works

The MCP server is available at:

```text theme={null}
https://app.beyondnoise.io/api/mcp
```

Your AI client connects to the endpoint, authenticates, and can then call any of the 8 available tools. All tool calls execute as your authenticated user.

## Connecting your AI client

### Claude Desktop

Add the following to your `claude_desktop_config.json`:

```json theme={null}
{
  "mcpServers": {
    "beyondnoise": {
      "url": "https://app.beyondnoise.io/api/mcp"
    }
  }
}
```

### Cursor

Add the server in **Settings > MCP** or edit `.cursor/mcp.json`:

```json theme={null}
{
  "mcpServers": {
    "beyondnoise": {
      "url": "https://app.beyondnoise.io/api/mcp"
    }
  }
}
```

## Available tools

| Tool                         | Description                                        |
| ---------------------------- | -------------------------------------------------- |
| `list_workspaces`            | List all workspaces you have access to             |
| `run_research`               | Run an ad-hoc research query                       |
| `read_research`              | Read a saved research result by ID                 |
| `create_continuous_research` | Create a scheduled continuous research topic       |
| `update_continuous_research` | Update an existing continuous research topic       |
| `read_continuous_research`   | Read a continuous research topic by ID             |
| `read_brand_context`         | Read the brand context for a workspace             |
| `update_brand_context`       | Create or update the brand context for a workspace |

<Tip>
  Call `list_workspaces` first to get valid `workspaceId` values for the other tools.
</Tip>

### list\_workspaces

Lists all workspaces the authenticated user can access. Returns `id`, `name`, `slug`, `website`, and `role` for each workspace.

**Parameters:** none

### run\_research

Runs an ad-hoc research query for a workspace. Returns the generated content, citations, and the result ID.

| Parameter             | Type      | Required | Description                                                         |
| --------------------- | --------- | -------- | ------------------------------------------------------------------- |
| `workspaceId`         | UUID      | Yes      | The workspace to run research in                                    |
| `content`             | string    | Yes      | The research question                                               |
| `level`               | enum      | No       | Research depth: `auto`, `minimal`, `low`, `medium`, `high`, `xhigh` |
| `includeBrandContext` | boolean   | No       | Include the workspace's brand context                               |
| `dataSources`         | string\[] | No       | Specific data sources to query                                      |
| `resultId`            | UUID      | No       | Pass with `previousRunId` to continue a thread                      |
| `previousRunId`       | string    | No       | Previous run ID for thread continuation                             |

### read\_research

Reads a saved ad-hoc research result by ID, including its message thread.

| Parameter  | Type | Required | Description            |
| ---------- | ---- | -------- | ---------------------- |
| `resultId` | UUID | Yes      | The research result ID |

### create\_continuous\_research

Creates a continuous research topic with a schedule and runs the first research immediately. Returns content, citations, and the topic ID.

| Parameter             | Type      | Required | Description                                                         |
| --------------------- | --------- | -------- | ------------------------------------------------------------------- |
| `workspaceId`         | UUID      | Yes      | The workspace to create the topic in                                |
| `content`             | string    | Yes      | The research question                                               |
| `level`               | enum      | No       | Research depth: `auto`, `minimal`, `low`, `medium`, `high`, `xhigh` |
| `schedule`            | enum      | No       | How often to run: `daily`, `weekly`, `monthly`                      |
| `includeBrandContext` | boolean   | No       | Include the workspace's brand context                               |
| `dataSources`         | string\[] | No       | Specific data sources to query                                      |

### update\_continuous\_research

Updates an existing continuous research topic. Only provided fields are updated.

| Parameter     | Type      | Required | Description                                                         |
| ------------- | --------- | -------- | ------------------------------------------------------------------- |
| `topicId`     | UUID      | Yes      | The continuous research topic ID                                    |
| `prompt`      | string    | No       | New research question                                               |
| `level`       | enum      | No       | Research depth: `auto`, `minimal`, `low`, `medium`, `high`, `xhigh` |
| `schedule`    | enum      | No       | How often to run: `daily`, `weekly`, `monthly`                      |
| `isActive`    | boolean   | No       | Activate or pause the topic                                         |
| `dataSources` | string\[] | No       | Specific data sources to query                                      |

### read\_continuous\_research

Reads a continuous research topic by ID, including its schedule and the latest result.

| Parameter | Type | Required | Description                      |
| --------- | ---- | -------- | -------------------------------- |
| `topicId` | UUID | Yes      | The continuous research topic ID |

### read\_brand\_context

Reads the brand context prompt for a workspace.

| Parameter     | Type | Required | Description      |
| ------------- | ---- | -------- | ---------------- |
| `workspaceId` | UUID | Yes      | The workspace ID |

### update\_brand\_context

Creates or updates the brand context prompt for a workspace.

| Parameter     | Type   | Required | Description            |
| ------------- | ------ | -------- | ---------------------- |
| `workspaceId` | UUID   | Yes      | The workspace ID       |
| `prompt`      | string | Yes      | The brand context text |

## Example usage

Once connected, you can ask your AI client things like:

* "List my workspaces"
* "Run a research on my top 3 competitors' pricing strategies"
* "Create a weekly continuous research tracking AI features in my competitors' products"
* "Read the brand context for my workspace"
* "Update my brand context to emphasize our focus on enterprise customers"

The AI client will automatically select the right tool and parameters based on your request.