FoundryNet Scrape

x402-gated web extraction gateway for AI agents. Fetch any page and get clean content back — text, markdown, or structured JSON. No API keys, no accounts.

Part of the FoundryNet Data Network. Pay-per-call in USDC (x402) or bypass with an fnet_ Forge key. Also exposed as an MCP server for Smithery/Glama/Claude.

Tools / Endpoints

Tool / Route	Price	Description
`extract` — `POST /v1/extract`	$0.01	`{url, format?: "text"\|"markdown"\|"json"}` → extracted main content
`extract_batch` — `POST /v1/extract/batch`	$0.008	`{urls: string[]}` (max 20) → array of extracted content (volume discount)

Backends: Mozilla Readability (article extraction) + Turndown (HTML→markdown) + Cheerio (fallback) — all free, no API key. Boilerplate (scripts, nav, ads) is stripped automatically.

MCP

Streamable HTTP endpoint: https://foundrynet-scrape-production.up.railway.app/mcp

claude mcp add --transport http foundrynet-scrape https://foundrynet-scrape-production.up.railway.app/mcp

Tools list without auth (discoverable). Pass an fnet_ key as api_key (or Authorization: Bearer) to bypass payment; otherwise the tool returns an x402 payment_required object.

Payment (x402)

Standard x402 v2 on Solana mainnet USDC. Discovery: GET /x402, GET /.well-known/x402, per-route 402 at GET /x402/{route}. Validated on 402 Index.

The highest-frequency tool category — every research, content, and data agent needs to read the web.