SmartFetch

SmartFetch is a webpage reader and fetch service for AI agents. It can scrape a public URL, extract clean text, convert a website to Markdown, preserve links and metadata, and use automatic browser rendering for JavaScript-heavy pages.

Each paid HTTP or MCP execution uses x402 exact payments at a default price of $0.005. Discovery, health, metadata, MCP initialize, and MCP tools/list remain free.

Public endpoints

• GET /meta — machine-readable metadata
• GET /.well-known/x402 — community x402 service manifest
• GET /openapi.json — OpenAPI 3.1
• GET /llms.txt — agent discovery summary
• POST /mcp — MCP Streamable HTTP
• POST https://smartfetch-production-ea53.up.railway.app/fetch — paid webpage retrieval

MCP tools

fetch_webpage

Return the full SmartFetch text, Markdown, links, and metadata result.

webpage_to_markdown

Return clean Markdown and core retrieval metadata without duplicate full text.

extract_webpage_text

Return clean readable text and core retrieval metadata without Markdown.

render_webpage

Force browser rendering and return the full SmartFetch result.

Paying client examples

The examples list tools for free, enforce a $0.005 maximum payment, then perform the x402 challenge, payment, retry, and settlement flow. Running them can spend real Base-mainnet USDC.

• Python paying MCP client
• TypeScript paying MCP client

HTTP request example

{"url":"https://example.com/","max_chars":20000,"force_browser":false}

Source and deployment documentation: https://github.com/Friezaaaa/smartfetch.