# @clawfetch/mcp

[![AllMCPs Verified](https://allmcps.com/api/badge/mcp-6)](https://allmcps.com/mcp/mcp-6?verify=d2c7d26f-b35a-4507-acd2-c29cca9186e4)

MCP (Model Context Protocol) server for [ClawFetch](https://clawfetch.ai) — Web Intelligence API for AI Agents.

Gives any MCP-compatible client (Claude Desktop, Cursor, Windsurf, OpenClaw, etc.) instant access to ClawFetch's web intelligence tools: fetch pages as markdown, render JS-heavy SPAs, extract structured data from 17+ site types, research topics multi-source, enrich companies from a domain, and check domain availability.

Payments are handled automatically via the [x402 protocol](https://www.x402.org/) — USDC on Base. No API keys, no billing dashboards.

## Access and payment

The public hosted MCP endpoint is retired and returns HTTP 410 Gone. Run the MCP locally using the setup below, or call https://api.clawfetch.ai directly. Paid tool calls use your own USDC-funded wallet on Base; ClawFetch does not subsidize requests.

## 💻 Self-Hosted Mode

Run your own instance with full control over the payment wallet.

### Install

```bash
npm install -g @clawfetch/mcp
```

### Configure (stdio — for local MCP clients)

Add to your MCP client config (e.g., Claude Desktop):

```json
{
  "mcpServers": {
    "clawfetch": {
      "command": "clawfetch-mcp",
      "env": {
        "CLAWFETCH_PRIVATE_KEY": "0xYOUR_ETHEREUM_PRIVATE_KEY"
      }
    }
  }
}
```

The private key signs x402 micropayments (USDC on Base). Fund the corresponding address with USDC on Base mainnet.

### Run as HTTP Server

Expose the MCP server over HTTP for remote clients:

```bash
CLAWFETCH_PRIVATE_KEY=0x... CLAWFETCH_TRANSPORT=http CLAWFETCH_PORT=3001 clawfetch-mcp
```

This starts a server at `http://localhost:3001/mcp` using the Streamable HTTP transport (MCP spec compliant). By default it binds to `127.0.0.1` and accepts only loopback `Host` headers, which blocks DNS-rebinding attacks from web pages. To serve remote clients, set `CLAWFETCH_HOST=0.0.0.0` and list your public hostnames in `CLAWFETCH_ALLOWED_HOSTS`.

#### HTTP Endpoints

| Endpoint | Method | Description |
|----------|--------|-------------|
| `/mcp` | POST | Initialize session — returns `Mcp-Session-Id` header |
| `/mcp` | GET | SSE stream for server-to-client notifications (requires session ID) |
| `/mcp` | DELETE | Close a session |
| `/health` | GET | Server health check (JSON) |

#### Session Flow

1. **Initialize:** `POST /mcp` with JSON-RPC `initialize` request → get `Mcp-Session-Id` header
2. **Call tools:** `POST /mcp` with `Mcp-Session-Id` header + JSON-RPC `tools/call` request
3. **Stream:** `GET /mcp?sessionId=...` for SSE notifications (optional)
4. **Close:** `DELETE /mcp` with `Mcp-Session-Id` header

## Tools

| Tool | Description | Cost |
|------|-------------|------|
| `fetch_url` | Fetch any URL as clean markdown | $0.001 |
| `render_page` | Render JS-heavy pages (SPAs, React apps) | $0.005 |
| `extract_data` | Extract structured data from 17+ site types | $0.008 |
| `research_topic` | Multi-source research with citations | $0.02 |
| `check_domains` | Check domain name availability | $0.003 |
| `suggest_domains` | Generate available domain suggestions | $0.003 |
| `parse_document` | Parse office docs (docx/pptx/xlsx/pdf/…) to markdown | $0.005 |
| `enrich_company` | Company profile from a domain: socials, emails, tech stack, key pages | $0.02 |
| `list_extractors` | List supported structured data extractors | $0.001 |
| `health_check` | Check API health status | Free |
| `wallet_info` | Show payment wallet address | Free |

### Supported Extractors

GitHub repos/profiles, npm packages, PyPI packages, Twitter/X profiles, LinkedIn profiles/companies, YouTube videos/channels, Product Hunt, Hacker News, Reddit, Crunchbase, Wikipedia, and more.

Use `list_extractors` to see all available types with their schema.

## Environment Variables

| Variable | Required | Default | Description |
|----------|----------|---------|-------------|
| `CLAWFETCH_PRIVATE_KEY` | Yes* | — | Ethereum private key (0x-prefixed) for x402 payments |
| `CLAWFETCH_BASE_URL` | No | `https://api.clawfetch.ai` | API base URL |
| `CLAWFETCH_TRANSPORT` | No | `stdio` | Transport: `stdio` or `http` |
| `CLAWFETCH_HOST` | No | `127.0.0.1` | HTTP server bind address |
| `CLAWFETCH_ALLOWED_HOSTS` | No | `localhost,127.0.0.1,[::1]` | Comma-separated `Host` header allowlist for the HTTP transport (DNS-rebinding protection) |
| `CLAWFETCH_PORT` | No | `3001` | HTTP server port |
| `CLAWFETCH_TIMEOUT_MS` | No | `60000` | Request timeout in ms |
| `CLAWFETCH_DEBUG` | No | `false` | Enable debug logging (`1` or `true`) |

\* Required for tool calls. Without it the server still starts and lists its tools, resources and prompts (so registries can inspect it), and every call returns setup instructions. Keep this key in your local configuration; never submit it to the ClawFetch website.

## Development

```bash
git clone https://github.com/clawfetch/clawfetch-mcp.git
cd clawfetch-mcp
npm install
npm run build
npm test
```

## License

MIT