MCP Server

Overview

Reqline includes a built-in Model Context Protocol (MCP) server that starts automatically with the desktop app. AI agents can use it to manage collections, execute requests, and work with environments programmatically.

The MCP server is available over Streamable HTTP at:

http://127.0.0.1:3218/mcp

Connecting AI clients

Claude Desktop / Cursor

Add to your MCP configuration:

{
  "mcpServers": {
    "reqline": {
      "url": "http://127.0.0.1:3218/mcp"
    }
  }
}

Warp

Add to ~/.warp/.mcp.json (global) or .warp/.mcp.json (per-project):

{
  "mcpServers": {
    "reqline": {
      "url": "http://127.0.0.1:3218/mcp"
    }
  }
}

Alternatively, use stdio mode (launches Reqline headlessly):

{
  "mcpServers": {
    "reqline": {
      "command": "pnpm",
      "args": ["start:mcp"]
    }
  }
}

Available tools

The MCP server exposes 20 tools:

Collections

Tool	Description
list_collections	List all saved collections
get_collection	Fetch one collection by ID
create_collection	Create a new collection
update_collection	Update an existing collection
delete_collection	Delete a collection and its requests

Requests

Tool	Description
list_requests	List saved requests (optionally by collection)
get_request	Fetch one request by ID
create_request	Create a new saved request
update_request	Update an existing request
delete_request	Delete a saved request

Environments

Tool	Description
list_environments	List all environments
get_environment	Fetch one environment by ID
create_environment	Create a new environment
update_environment	Update an existing environment
delete_environment	Delete an environment
get_active_environment	Get the currently active environment
set_active_environment	Activate an environment by ID
clear_active_environment	Deactivate the current environment

Execution

Tool	Description
execute_saved_request	Execute a saved request by ID using the active environment
execute_request	Execute an ad-hoc request payload

Both execution tools record the result in history, just like sending from the UI.

Resources

The MCP server also exposes two read-only resources:

• reqline://collections — Collection metadata (names, IDs, timestamps)
• reqline://requests — Request metadata (names, methods, URLs, IDs) without sensitive data

Configuration

Environment variable	Default	Description
REQLINE_MCP_PORT	3218	MCP HTTP server port
REQLINE_LOCAL_API_PORT	3217	Local API server port
REQLINE_LOCAL_API_TOKEN	(none)	Optional auth token for the local API

If the configured port is occupied, the server automatically falls back to a random available port.