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

# Server

> Connect a remote MCP (Model Context Protocol) server to the workflow AI agent node

Connect a remote MCP (Model Context Protocol) server to the workflow AI agent node. The agent discovers the server's tools and can call them during a run. Supports streamable HTTP servers with optional custom headers for authentication. Can only be used as a workflow AI agent tool, not as a standalone node or main-agent tool.

|                     |              |
| ------------------- | ------------ |
| **App**             | MCP          |
| **Operation ID**    | `mcp_server` |
| **Type**            | Action       |
| **Connection**      | None         |
| **Credits per run** | Free         |

## Inputs

| Field           | Type            | Required | Default | Description                                                                                                |
| --------------- | --------------- | -------- | ------- | ---------------------------------------------------------------------------------------------------------- |
| `url`           | string          | Yes      | -       | Streamable HTTP endpoint of the MCP server, e.g. `https://example.com/mcp.` Only https URLs are supported. |
| `headers`       | object          | No       | -       | Optional HTTP headers sent with every request to the MCP server, e.g. \{'Authorization': 'Bearer ...'}     |
| `allowed_tools` | array of string | No       | -       | Optional list of tool names the agent is allowed to use from this server. Leave empty to allow all tools.  |

## Output

**Type**: `None`

MCP tool results are returned to the AI agent, not as node output.

**Fields**: dynamic (depend on the inputs)