# Legible

> The files AI agents look for, checked on any site. With the real file from whoever does it best on our category leaderboards.

Legible runs 28 agent-readiness checks on any domain, and 3 more that count only for email APIs. It has 11 category leaderboards, 150 companies in all, each board from one dated scan: Email APIs (15), AI agent platforms (13), Payments (15), Hosting and deploy (12), Observability (15), Databases and caching (16), Auth (14), File storage (15), Background jobs (13), Feature flags (12), Real-time (10). Free, no account.

## Pages
- [The leaderboards](https://leaderfile.vercel.app/leaderboards): all 11 boards and every company ranked, their scores and what the leaders ship
- [How we check](https://leaderfile.vercel.app/method): every check, how it's scored and how our scanner behaves
- [Best practices](https://leaderfile.vercel.app/best-practices): how to do each check, with the real llms.txt, OpenAPI, MCP and install pages the boards ship
- [Compare](https://leaderfile.vercel.app/compare): one domain against its rivals, check by check, with the real file from whoever passes
- [The Drop, issue 3](https://leaderfile.vercel.app/issue/3): The first weekly Drop: what moved across 13 AI agent platforms since 3 Nov, why the email board stayed put, and one real file worth copying. (newest issue at https://leaderfile.vercel.app/drop; feed at https://leaderfile.vercel.app/feed.xml)
- [Issue 2](https://leaderfile.vercel.app/issue/2): What changed across 15 email APIs between our 29 Oct and 2 Nov scans, and the best file we found.
- [Issue 1](https://leaderfile.vercel.app/issue/1): Every agent-readiness check, and which of 15 email APIs pass it, from our 29 Oct 2026 scan.
- [Our picks](https://leaderfile.vercel.app/picks): hand-picked open-source tools and agent skills that help pass our checks, with GitHub's numbers, read weekly
- [Chrome extension](https://leaderfile.vercel.app/extension): press it on any site to scan the one you're on
- [Pricing](https://leaderfile.vercel.app/pricing): free

## API
- [Developer docs](https://leaderfile.vercel.app/docs): the command line, the scan API, the MCP server, the agent skill and a GitHub Actions job
- [Command line](https://leaderfile.vercel.app/docs#cli): npx legible-cli <domain>, or npm install -g legible-cli and then legible <domain>; --json for scripts, --min-score <n> for CI, and npx legible-cli skill adds the agent skill to a project
- [Scan API](https://leaderfile.vercel.app/docs): POST /api/scan with a domain or a board company
- [OpenAPI spec](https://leaderfile.vercel.app/openapi.json): the same API as OpenAPI 3.1
- [Legible's MCP server](https://leaderfile.vercel.app/mcp): scan_site, get_checks and get_board, for an agent to call from your editor
- [Agent skill](https://leaderfile.vercel.app/skill.md): a SKILL.md for a coding agent to scan a live site and write its own version of each file
- [Skills for each practice](https://leaderfile.vercel.app/skills): one SKILL.md per practice (llms.txt, discovery, docs, an API spec, MCP, packages) for building them; the list as JSON at https://leaderfile.vercel.app/skills/index.json