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

# Connect your AI client

> Step-by-step setup for Claude, Claude Code, ChatGPT, Codex, Cursor, VS Code and any other MCP client.

Every client connects to the same endpoint:

```text theme={null}
https://api.evermuse.com/api/mcp
```

The first time your client connects, it opens the Evermuse sign-in page in your browser. Sign in, choose the workspace
the agent may access, and click **Allow access**. The client then handles tokens and refreshes automatically. For what
happens during sign-in, see [Authentication](/mcp/authentication).

<Tip>
  You can also find these instructions inside Evermuse under **Settings → MCP**, with the endpoint ready to copy.
</Tip>

## Claude

<Tabs>
  <Tab title="Claude (web, Desktop, mobile)">
    <Steps>
      <Step title="Open connector settings">
        In Claude, open **Settings → Connectors**.
      </Step>

      <Step title="Add a custom connector">
        Click **Add custom connector**, name it `Evermuse`, and paste `https://api.evermuse.com/api/mcp` as the URL.
        Leave the advanced OAuth fields empty: Claude registers itself with Evermuse automatically.
      </Step>

      <Step title="Connect">
        Click **Connect**, sign in to Evermuse, choose your workspace and click **Allow access**.
      </Step>

      <Step title="Try it">
        Start a new chat, make sure Evermuse is enabled in the tools menu, and ask:
        *"What are customers saying about onboarding?"*
      </Step>
    </Steps>

    <Note>
      On Claude Team and Enterprise plans, an organization owner may need to add the connector for the organization
      before members can connect it. Connectors added on claude.ai are available in Claude Desktop and the Claude mobile apps.
    </Note>
  </Tab>

  <Tab title="Claude Code">
    Add the server from your terminal:

    ```bash theme={null}
    claude mcp add --transport http evermuse https://api.evermuse.com/api/mcp
    ```

    Then start Claude Code, run `/mcp`, select **evermuse** and choose **Authenticate**. Your browser opens the
    Evermuse sign-in page. After you approve, return to the terminal.

    Add `--scope user` to make Evermuse available in every project, or `--scope project` to share it with your team
    through a checked-in `.mcp.json`.

    Evermuse skills appear as slash commands, for example `/mcp__evermuse__create-prd`. See
    [Skills & prompts](/mcp/skills-and-prompts).
  </Tab>
</Tabs>

## ChatGPT

<Steps>
  <Step title="Open connector settings">
    In ChatGPT, open **Settings → Apps & Connectors** (called **Connectors** in some plans). Adding a custom connector
    is available on plans that support custom MCP connectors, such as Pro, Business and Enterprise. Depending on your
    plan, you may need to turn on **Developer mode** under **Advanced settings** first, and workspace admins may need to
    allow custom connectors.
  </Step>

  <Step title="Create the connector">
    Click **Create** (or **Add custom connector**), name it `Evermuse`, paste `https://api.evermuse.com/api/mcp` as the
    MCP server URL, and choose **OAuth** as the authentication method.
  </Step>

  <Step title="Sign in">
    Sign in to Evermuse, choose your workspace and click **Allow access**.
  </Step>

  <Step title="Use it in a chat">
    In a new chat, enable Evermuse from the tools menu and ask a product question.
  </Step>
</Steps>

## Codex

Add Evermuse to `~/.codex/config.toml`:

```toml ~/.codex/config.toml theme={null}
[mcp_servers.evermuse]
url = "https://api.evermuse.com/api/mcp"
```

Then sign in over OAuth:

```bash theme={null}
codex mcp login evermuse
```

## Cursor

<Steps>
  <Step title="Add the server">
    Open **Cursor Settings → MCP → Add new MCP server**, or add Evermuse to `~/.cursor/mcp.json` (all projects) or
    `.cursor/mcp.json` (one project):

    ```json mcp.json theme={null}
    {
      "mcpServers": {
        "evermuse": {
          "url": "https://api.evermuse.com/api/mcp"
        }
      }
    }
    ```
  </Step>

  <Step title="Sign in">
    In the MCP settings, click **Login** (or **Needs login**) next to Evermuse and complete the sign-in in your browser.
  </Step>
</Steps>

## VS Code

GitHub Copilot agent mode in VS Code supports remote MCP servers with OAuth. Add Evermuse to `.vscode/mcp.json` in your
workspace, or to your user `mcp.json` with the **MCP: Open User Configuration** command:

```json .vscode/mcp.json theme={null}
{
  "servers": {
    "evermuse": {
      "type": "http",
      "url": "https://api.evermuse.com/api/mcp"
    }
  }
}
```

Start the server from the `mcp.json` editor or the **MCP: List Servers** command, and approve the sign-in prompt.

## Other MCP clients

Any client that supports the Streamable HTTP transport and the MCP authorization specification works. Point it at the
endpoint and authenticate when prompted. The client discovers everything else automatically:

1. Its first request without a token receives `401 Unauthorized` with a `WWW-Authenticate` header that points to
   `https://api.evermuse.com/.well-known/oauth-protected-resource`.
2. That document names the authorization server, `https://api.evermuse.com`, whose metadata is at
   `/.well-known/oauth-authorization-server`.
3. The client registers itself (usually through Dynamic Client Registration), runs the authorization
   code flow with PKCE, and calls the server with `Authorization: Bearer <access_token>`.

Full protocol details are in [Authentication](/mcp/authentication).

For a client that only supports a fixed header and cannot run OAuth, use an
[API key](/mcp/authentication#api-keys) instead:

```json theme={null}
{
  "mcpServers": {
    "evermuse": {
      "url": "https://api.evermuse.com/api/mcp",
      "headers": { "x-api-key": "em_sk_your_api_key" }
    }
  }
}
```

### Test with MCP Inspector

To explore the server interactively, run the official
[MCP Inspector](https://modelcontextprotocol.io/docs/tools/inspector):

```bash theme={null}
npx @modelcontextprotocol/inspector
```

Choose the **Streamable HTTP** transport, enter `https://api.evermuse.com/api/mcp`, and use **Open Auth Settings → Quick
OAuth Flow** to sign in. Then list tools and call `get_products` to confirm the connection.

## Verify the connection

Ask your agent: *"Which Evermuse products can you see?"* It should call `get_products` and list the products in the
workspace you authorized. If it doesn't, see [Troubleshooting](/mcp/troubleshooting).

## Switching workspaces

Each connection is bound to one workspace, chosen when you approve access. To use a different workspace, disconnect
Evermuse in your client and connect again, choosing the other workspace on the consent screen. See
[Revoking access](/mcp/authentication#revoking-access).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.