OpenClaw MCP Server
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@OpenClaw MCP Serverask my assistant about the status of the deployment"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
OpenClaw MCP Server ๐ฐ
Expose your OpenClaw assistant to Claude Desktop (or any MCP client) via Streamable HTTP transport.
This allows you to interact with your OpenClaw assistant from Claude Desktop, effectively bridging two AI systems.
โจ Features
Streamable HTTP transport โ proper MCP spec implementation
SSE streaming โ real-time responses
API Key authentication โ secure remote access
Easy installation โ one script to set everything up
Configurable โ customize bot name, session label, port
systemd integration โ auto-start on boot
Related MCP server: OpenClaw MCP Server
๐ Quick Install
On the server where OpenClaw is running:
# Clone the repo
git clone https://github.com/rodgco/openclaw-mcp-server.git
cd openclaw-mcp-server
# Run installer
./install.shThe installer will:
โ Check for Node.js 18+ and OpenClaw
๐ฆ Install dependencies
๐ Generate an API key
โ๏ธ Ask for your bot name and session label
๐ Optionally create a systemd service
๐ Output the Claude Desktop configuration
๐ Manual Installation
If you prefer manual setup:
# Clone and install
git clone https://github.com/rodgco/openclaw-mcp-server.git
cd openclaw-mcp-server
npm install
# Configure
cp .env.example .env
# Edit .env with your settings
# Generate API key
openssl rand -base64 32
# Run
npm startโ๏ธ Configuration
Edit .env file:
# Name of your assistant/bot
BOT_NAME="MyAssistant"
# OpenClaw session label to connect to
OPENCLAW_SESSION_LABEL="main"
# API Key (generate with: openssl rand -base64 32)
MCP_SERVER_API_KEY="your-secret-key-here"
# Port to listen on
PORT=3721
# Bind address (0.0.0.0 for all interfaces)
BIND_ADDRESS="0.0.0.0"
# OpenClaw workspace path
OPENCLAW_WORKSPACE="${HOME}/.openclaw/workspace"๐ฅ๏ธ Claude Desktop Configuration
Add to your claude_desktop_config.json:
{
"mcpServers": {
"myassistant": {
"transport": "streamable-http",
"url": "http://YOUR_SERVER_IP:3721/mcp",
"headers": {
"Authorization": "Bearer YOUR_API_KEY"
}
}
}
}Config file locations:
macOS:
~/Library/Application Support/Claude/claude_desktop_config.jsonWindows:
%APPDATA%\Claude\claude_desktop_config.jsonLinux:
~/.config/Claude/claude_desktop_config.json
๐ ๏ธ Available Tools
Once connected, Claude Desktop can use these tools:
ask
Send a message or question to your OpenClaw assistant.
memory_search
Search the assistant's long-term memory (MEMORY.md).
sessions_status
Check OpenClaw session status.
๐ก How It Works
Claude Desktop (your computer)
โ๏ธ HTTP POST/GET + SSE
http://your-server:3721/mcp
โ๏ธ
openclaw-mcp-server
โ๏ธ openclaw sessions send
OpenClaw Assistant (remote server)Claude Desktop sends JSON-RPC requests via HTTP POST
MCP server forwards messages to OpenClaw session
Responses stream back via Server-Sent Events (SSE)
๐ง systemd Service
If you installed with the script and chose systemd:
# Check status
sudo systemctl status openclaw-mcp-server
# View logs
sudo journalctl -u openclaw-mcp-server -f
# Restart
sudo systemctl restart openclaw-mcp-server๐ Security
API Key required for all MCP requests
Tailscale recommended for secure remote access
Never expose port 3721 to the public internet without additional protection
Consider using HTTPS in production (reverse proxy with nginx/caddy)
๐งช Testing
# Health check
curl http://localhost:3721/health
# Server info
curl http://localhost:3721/
# Test MCP endpoint (requires auth)
curl -X POST \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-H "Accept: text/event-stream" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' \
http://localhost:3721/mcp๐๏ธ Uninstall
./uninstall.shOr manually:
sudo systemctl stop openclaw-mcp-server
sudo systemctl disable openclaw-mcp-server
sudo rm /etc/systemd/system/openclaw-mcp-server.service
rm -rf ~/.openclaw/mcp-server๐ MCP Specification
This server implements the Streamable HTTP transport from the Model Context Protocol specification.
๐ค Contributing
Pull requests welcome! Please open an issue first to discuss changes.
๐ License
MIT
๐ Links
This server cannot be deployed
Maintenance
Related MCP Connectors
- QuallaaOAuthcom.quallaa
Talk to your public-facing AI from any MCP client โ Claude, ChatGPT, Cursor, Cline, Windsurf.
Deploy, monitor, and manage your OpenClaw AI assistants via natural language.
Drop-in proxy keeping OpenAI Assistants API calls working past the August 26, 2026 sunset, plus a Th
Remote streamable-HTTP MCP server running on a single Cloudflare Worker. Your assistant gets live Airbnb, Amazon, Booking.com, Google Flights, Maps and Reddit data, social search on X, Instagram and TikTok, the Meta Ad Library, and image/video generation without any keys. Connect your own accounts to let it send WhatsApp or Telegram messages, work an IMAP inbox, manage Meta Ads campaigns and publish to X and LinkedIn. OAuth 2.1 with PKCE; stored credentials are AES-256-GCM encrypted.
Related MCP Servers
- FlicenseNot gradedqualityBmaintenanceExposes locally-built AI tools to Claude, enabling article summarization, promo generation, and semantic search over an Obsidian vault.-
- AlicenseNot gradedqualityAmaintenanceEnables Claude.ai and Claude Desktop to connect securely to self-hosted OpenClaw assistants via OAuth 2.1, with tools for chat, asynchronous task handling, multi-instance routing, and gateway health checks.92 npmMIT
- AlicenseNot gradedqualityAmaintenanceEnables Claude.ai and Claude Desktop to delegate tasks to self-hosted OpenClaw assistant gateways through a secure, OAuth 2.1-authenticated bridge. Supports synchronous chat, asynchronous task queuing and polling, health checks, and routing requests across multiple isolated instances.92 npmMIT
- AlicenseNot gradedqualityAmaintenanceBridges Claude.ai and Claude Desktop to self-hosted OpenClaw AI assistants through an OAuth 2.1-secured HTTP endpoint. Enables sending chat messages, tracking long-running tasks, checking gateway health, and routing requests across multiple isolated instances from a single connection.92 npmMIT