FREE API FOR AI AGENTS
Before an agent fetches a website, Front Door tells it the legitimate ways in: official API, feeds, MCP, OpenAPI, llms.txt, what robots.txt and Content Signals say for AI training versus search versus an agent acting for its user, whether pages are an empty JavaScript shell, whether automated access is challenged, and what the terms say. Every field is dated.
Facts as observed and as the site states them. Not permission, not legal advice. Front Door never lists ways around a site's protections.
GET https://money-lab.app/api/fd?domain=github.com GET https://money-lab.app/api/fd/route?domain=arxiv.org&task=bulk GET https://money-lab.app/api/fd/changes?domain=reddit.com MCP: front_door_card, front_door_best_route, front_door_changes on https://money-lab.app/api/mcp
Free: 50 cards a day per network with nothing attached; 500 with a free key. A fresh probe on demand is $0.005 over x402; the whole set is $1.
Hand-curated so far (2026-10-22): github.com, wikipedia.org, arxiv.org, reddit.com, stackoverflow.com, news.ycombinator.com, x.com, youtube.com, suno.com, elevenlabs.io, canva.com, gumroad.com. Every other domain gets the automatic probe; the most-asked domains are curated next.