mcpField
E
ExtractEdge

ExtractEdge

Remote
github.com/vasilicasijarvisBrowser Automation0Updated 6w ago
Repository

About ExtractEdge

ExtractEdge is a web-page extraction service for AI agents that converts page content into clean markdown. It offers an MCP endpoint, a capped free demo, and paid extraction through x402.

Key features

  1. 01Main-content detection
  2. 02HTML to markdown conversion
  3. 03JSON and binary response handling
  4. 04Redirects and request limits
  5. 05Free capped demo extraction
  6. 06x402 pay-per-use extraction

Use cases

  1. 01Convert a web page into markdown for an agent
  2. 02Extract article content without page navigation elements
  3. 03Try limited extraction without payment
  4. 04Request full extraction with x402

MCP support

Its Streamable HTTP MCP endpoint exposes an `extract` tool that accepts a URL and can return capped demo markdown.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://extractedge.lastunmihai1986.workers.dev/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

Tools

Details

From the project README

ExtractEdge

Firecrawl-lite: turn any web page into clean LLM-ready markdown. Fast, cheap, crypto-native. Live: https://extractedge.lastunmihai1986.workers.dev

Why

Firecrawl starts at $16/mo for ~1k credits and requires signup + API key. ExtractEdge is $0.001/extract pay-per-use over x402 (USDC on Base mainnet), no signup, no API key — designed for AI agents.

Quick start (free demo)

curl -X POST "https://extractedge.lastunmihai1986.workers.dev/extract?demo=1" \
  -H "content-type: application/json" \
  -d '{"url":"https://example.com"}'

Free demo returns up to 3,000 chars of the page as markdown. No payment needed.

Paid extraction (x402, USDC on Base)

  1. POST /extract {"url":"https://target.page"} → 402 with PAYMENT-REQUIRED header (base64 JSON)
  2. Sign an EIP-3009 transferWithAuthorization for 1000 atomic units (0.001 USDC) on Base
    • asset: 0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913 (USDC)
    • payTo: 0xa1b8be63bde77ddc65d9ffe8a21a2a5f2c9ca6a8
  3. Retry with the PAYMENT-SIGNATURE header (base64 PaymentPayload JSON) → 200 with full markdown + PAYMENT-RESPONSE receipt

Python client example: see scripts/e2e_paid2.py (eth_account EIP-712 signing, ~40 lines).

MCP server

Streamable HTTP MCP endpoint (JSON-RPC 2.0):

https://extractedge.lastunmihai1986.workers.dev/mcp

REST endpoints

EndpointDescription
POST /extractPaid extraction (x402) or free demo (?demo=1 / X-ExtractEdge-Demo header)
POST /mcpMCP JSON-RPC (initialize / tools/list / tools/call / ping)
GET /.well-known/x402.jsonPayment discovery
GET /.well-known/agent-registration.jsonERC-8004 agent registration file
GET /Service info

Response shape

{
  "success": true,
  "format": "markdown",
  "content": "# Page title\n\nClean markdown...",
  "textContent": "plain text variant",
  "metadata": {
    "title": "Page title", "description": "meta description",
    "source": "https://final.url/", "statusCode": 200,
    "contentLength": 48211, "fetchMs": 731,
    "links": ["https://..."]
  },
  "payment": {"payer": "0x...", "transaction": "0x...", "amount": "1000"}
}

Features

  • Main-content detection (scores article/main/div candidates by text density, drops nav/footer/ads)
  • HTML → markdown: headings, bold/italic, links (reference-style), images, ordered/unordered lists (nested), code blocks, blockquotes, tables (GFM), hr
  • Entity decoding (named + numeric), script/style/svg stripped
  • JSON endpoints returned pretty-printed; binary resources answered with a notice
  • Redirect following (≤5), 15s timeout, 3MB cap, per-IP demo rate limit (20/min)
  • CORS enabled; abuse control on demo lane only — paid lane unlimited

Architecture

Single Cloudflare Worker (stdlib-only, zero npm deps):

  • src/index.js — router + info endpoints
  • src/extract.js — fetch pipeline (timeout, redirects, content-type dispatch)
  • src/md.js — HTMLRewriter → DOM tree → markdown/text + main-content scoring
  • src/x402.js — x402 v2 gate: 402/PAYMENT-REQUIRED, EIP-3009 verify+settle via Coinbase CDP facilitator (Ed25519 JWT per request), Bazaar discovery declaration
  • src/mcp.js — MCP Streamable HTTP server (JSON-RPC 2.0)

Pricing comparison

FirecrawlExtractEdge
Entry price$16/mo (~1k credits)$0 pay-per-use
Per page~$0.016$0.001
Signup/API keyrequirednone
PaymentcardUSDC on Base (x402)
Agent-native (MCP)hostedMCP + x402 + Bazaar

Status

  • Worker live on extractedge.lastunmihai1986.workers.dev (global edge)
  • 3 real settled payments on Base mainnet (tx receipts verified on-chain)
  • MCP published on Smithery (admin-qnk4/extractedge)

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.