Any API Documentation,
ready for your agent.
apiCrawl is a community-powered catalogue of searchable documentation of any API out there, with Agent-First Design.
Just drop API documentation URLs here, scrape it and use it in your agent. For Free, for real, no tricks.
Browse what's already crawled
Popular APIs, structured and ready. Open the catalogue, connect to your agent — nothing to install.
Point it at any API reference URL
apiCrawl scrapes any API documentation intelligently whether it's structured or not. No config needed.
Crawl once.
Use everywhere.
The catalogue grows every time someone runs the OSS tool. Everyone benefits.
Anyone crawls it
Just paste any API docs URL — here or via the skill. It discovers the underlying spec, falls back to smart crawling, classifies content with an LLM, and structures everything into clean markdown. No install needed.
We store it
Structured markdown lands in the public catalogue — endpoints, parameters, auth instructions, TypeScript types. Indexed and searchable. If it's already there, no one needs to crawl it again.
You use it
Browse the catalogue. Plug any API into your agent via MCP or the apiCrawl skill. No installation. No account. No crawling. Your agent gets the full structured docs instantly.
Plug into your agents
Two ways to give your agent access to the entire catalogue. Pick whichever fits your stack.
MCP Server
Connect the apiCrawl MCP server to your agent. Search the catalogue, fetch endpoint details, get auth instructions, and pull structured docs for any API — all via tool calls. No install needed.
# Add to your MCP config
{"mcpServers": {
"apicrawl": {
"type": "http",
"url": "https://api.apicrawl.dev/mcp"
}}apiCrawl Skill
Install the apiCrawl skill once. It connects your agent to the entire catalogue — not just one API. Search by name, fetch endpoint details, or ask questions about any of the 1,240+ catalogued APIs.
# Install the skill
npx skills add flowstash/api-crawl-skill
# Then use it in your agent
/api-crawl apiId: [apiId]
Check the docs on how to …API docs — indexed and easily searchable
Every catalogue entry is parsed, classified, and structured — not raw HTML dumped to text.
Auth extracted
Authentication instructions are pulled out and summarised — API keys, OAuth flows, Bearer tokens — so your agent doesn't have to hunt through docs.
LLM Optimised
OpenAPI specs are verbose, hard to search, and burn through tokens. We autocompress schema objects to TypeScript definitions — up to 50% fewer tokens with zero loss of meaning.
Parse about anything
Automatically finds OpenAPI/Swagger specs, Postman collections, and Google Discovery docs. Falls back to smart HTML crawling — and picks up llms.txt indexes — when no spec exists.
Sections + BM25 index
Docs are split into structured sections and indexed with a context-aware BM25 index. Your agent can search semantically and drill into specifics — just like a developer would.
Built in the open.
Free forever.
The managed catalogue is free with no account needed. The crawler is Apache-2.0 open source — use it to add APIs to the catalogue, self-host everything, or run it locally against your own docs.
View on GitHubStart using the catalogue
Browse 1,240+ already-crawled APIs or add one that's missing. No account, no installation, no catch.
Free forever · No sign-up required · Open Source