Openclaw Mcp

Name: Openclaw Mcp
Author: freema

by freemacommunityv1.4.1

169

MCP server that bridges Claude Desktop or Claude.ai to an OpenClaw gateway, letting the client send chats, check gateway health, and manage tasks against one or more OpenClaw instances. It supports both synchronous and asynchronous operations with OAuth/CORS-protected access for remote use.

From the registry: MCP server bridging Claude.ai/Desktop with self-hosted OpenClaw via OAuth 2.1.

GitHub Repository Website

Install & configure

Prerequisites

required
Install Node.js 20 or newer
Node.js version 20 or higher is required to run the server via npx.
required
Run an OpenClaw gateway with HTTP API enabled
The server connects to an OpenClaw gateway. The gateway must be running and have the chat completions HTTP endpoint enabled.
https://github.com/openclaw/openclaw
required
Obtain an OpenClaw gateway token
You need a gateway token for the MCP server to authenticate to OpenClaw.

Install in your client

Paste into the AI you're chatting with — it will figure out the rest

Please install the `openclaw-mcp` MCP server into my current AI client (that's you).

Required prerequisites (do these first if not already done):
- **Install Node.js 20 or newer** — Node.js version 20 or higher is required to run the server via npx.
- **Run an OpenClaw gateway with HTTP API enabled** — The server connects to an OpenClaw gateway. The gateway must be running and have the chat completions HTTP endpoint enabled. (https://github.com/openclaw/openclaw)
- **Obtain an OpenClaw gateway token** — You need a gateway token for the MCP server to authenticate to OpenClaw.

Canonical MCP server config (stdio transport):
- command: `npx`
- args: ["openclaw-mcp"]
- required environment variables:
  - `OPENCLAW_URL`: Base URL of your OpenClaw gateway. (example: `<your-openclaw-url>`)
  - `OPENCLAW_GATEWAY_TOKEN`: Gateway token used to authenticate to OpenClaw. (example: `<your-gateway-token>`)
- optional environment variables:
  - `OPENCLAW_MODEL`: Model name to use when sending requests to OpenClaw. README examples use 'openclaw'. (example: `<your-openclaw-model>`)
  - `OPENCLAW_TIMEOUT_MS`: Request timeout in milliseconds. (example: `<your-timeout-ms>`)
  - `OPENCLAW_INSTANCES`: JSON array defining multiple OpenClaw gateway instances for multi-instance mode. When set, it can be used instead of single-instance URL/token settings. (example: `<your-json-instances-config>`)

Note: The simplest portable install path is stdio via `npx openclaw-mcp` for local MCP clients. The README also documents remote SSE mode (`npx openclaw-mcp --transport sse --port 3000`) for Claude.ai and Docker deployments. For SSE/remote mode, additional auth-related environment variables may be needed, including AUTH_ENABLED, MCP_CLIENT_ID, MCP_CLIENT_SECRET, MCP_ISSUER_URL, and CORS_ORIGINS.

Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.

Note: The simplest portable install path is stdio via `npx openclaw-mcp` for local MCP clients. The README also documents remote SSE mode (`npx openclaw-mcp --transport sse --port 3000`) for Claude.ai and Docker deployments. For SSE/remote mode, additional auth-related environment variables may be needed, including AUTH_ENABLED, MCP_CLIENT_ID, MCP_CLIENT_SECRET, MCP_ISSUER_URL, and CORS_ORIGINS.