OpenClaw MCP Server Guide: Connect Any API in Under 10 Minutes

The Model Context Protocol (MCP) is the secret sauce that makes OpenClaw so powerful. While skills provide procedural knowledge, MCP servers provide live connectivity to external data and services. In this guide, I'll show you how to connect any API to your OpenClaw agent in under 10 minutes.

Why MCP Matters for OpenClaw

Before MCP, connecting an agent to a new service meant writing custom tool definitions, handling authentication manually, and managing API responses. MCP standardizes this. Once you have an MCP server running, OpenClaw can automatically discover its tools, resources, and prompts without any additional configuration.

I use MCP servers daily to bridge the gap between my agent's reasoning and the real-world data living in my databases, CRMs, and custom internal tools.

The Anatomy of an MCP Connection

An MCP connection consists of three parts:

1. The Server: A process (Node.js, Python, etc.) that implements the MCP spec.
2. The Transport: Usually standard input/output (stdio) or HTTP/SSE.
3. The Client: OpenClaw, which connects to the server and executes tools.

Step 1: Setting Up Your First MCP Server

The fastest way to get started is using a community-built server or a simple Node.js implementation. Let's build a basic "Weather MCP" server as an example.

# Create a new directory
mkdir my-mcp-server && cd my-mcp-server

# Initialize npm
npm init -y

# Install the MCP SDK
npm install @modelcontextprotocol/sdk

Now, create a file named index.js with a basic tool implementation:

import { Server } from "@modelcontextprotocol/sdk/server/index.js";
import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";
import { CallToolRequestSchema, ListToolsRequestSchema } from "@modelcontextprotocol/sdk/types.js";

const server = new Server({
  name: "my-weather-server",
  version: "1.0.0",
}, {
  capabilities: {
    tools: {},
  },
});

server.setRequestHandler(ListToolsRequestSchema, async () => ({
  tools: [{
    name: "get_weather",
    description: "Get current weather for a city",
    inputSchema: {
      type: "object",
      properties: {
        city: { type: "string" },
      },
      required: ["city"],
    },
  }],
}));

server.setRequestHandler(CallToolRequestSchema, async (request) => {
  if (request.params.name === "get_weather") {
    const { city } = request.params.arguments;
    // In a real app, you'd fetch from an API here
    return {
      content: [{ type: "text", text: `The weather in ${city} is sunny, 72°F.` }],
    };
  }
  throw new Error("Tool not found");
});

const transport = new StdioServerTransport();
await server.connect(transport);

Step 2: Configuring OpenClaw

To tell OpenClaw about your new server, you need to add it to your openclaw.jsonconfiguration file. This file is typically located at ~/.openclaw/openclaw.json.

{
  "mcpServers": {
    "weather": {
      "command": "node",
      "args": ["/absolute/path/to/my-mcp-server/index.js"]
    }
  }
}

Once you save this file, OpenClaw will automatically launch the server whenever it starts up and register the get_weather tool.

Step 3: Usage in Conversations

You don't need to do anything special to use the tool. Just ask your agent!

"What's the weather like in San Francisco?"

OpenClaw will see the get_weather tool in its available toolkit, see that it matches your intent, and call the MCP server to get the data.

Advanced Patterns: Resources and Prompts

MCP is more than just tools. You can also expose Resources (static or dynamic files/data) and Prompts (pre-defined templates).

• Resources: Great for giving the agent access to logs, documentation, or database snapshots.
• Prompts: Useful for standardizing how the agent handles specific tasks like code reviews or bug reports.

Troubleshooting Common Issues

If your MCP server isn't appearing in OpenClaw, check these three things:

1. Absolute Paths: Always use absolute paths in openclaw.json. OpenClaw might run from a different working directory.
2. Environment Variables: If your server needs API keys, ensure they are passed in the env section of the config.
3. Logs: Check the OpenClaw logs (usually via openclaw status) to see if the server failed to start.

Frequently Asked Questions

Internal Links

Ready to dive deeper? Check out these related guides:

• Building Custom Skills: A Complete Guide
• Mastering Browser Automation with OpenClaw
• Security Hardening Guide for Production Agents