MCP Setup Guide
How to Set Up Fetch MCP with ClawdBot
Turn any URL into clean Markdown for rapid ingestion.
Overview
- Category: Web
- Difficulty: Easy
- Best for: Turn any URL into clean Markdown for rapid ingestion.
Quick Start
If your provider offers a hosted MCP endpoint, use a URL-based config:
Hosted MCP Config json
{
"fetch": {
"url": "https://YOUR_FETCH_MCP_ENDPOINT"
}
} For local or self-hosted servers, use a CLI bridge and keep secrets in environment variables:
Local CLI Config json
{
"fetch": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://YOUR_FETCH_MCP_ENDPOINT"],
"env": {
"FETCH_API_KEY": "YOUR_FETCH_API_KEY"
}
}
} Common Pitfalls
- When to use Fetch vs Puppeteer.
- Store credentials in environment variables, not in config files.
- Limit scopes to the minimum required for your workflows.
- Verify the MCP is running with
clawdbot statusbefore testing.
Example Prompts
- Search the web for the latest MCP benchmarks and summarize the top 3 sources.
- Find a reliable source explaining Model Context Protocol.
- Collect three recent articles about AI tooling and summarize them.