Audit any website for AI Agent Readiness — the Lighthouse for the Agent Era.
Measure how well your application serves AI agents across five dimensions: Discovery, Identity, Auth & Access, Integration, and UX.
Built on the ora Agent Readiness framework.
# Install dependencies
pnpm install
# Start API (port 3000)
pnpm --filter api dev
# Start Web UI (port 3001)
pnpm --filter web dev
# Run CLI audit
node packages/cli/dist/cli.js vercel.com| Package | Description |
|---|---|
@agent-lighthouse/core |
Audit engine — scores any URL across 5 dimensions |
agent-lighthouse |
CLI tool — agent-lighthouse <url> |
apps/api |
Hono REST API — GET /api/audit?url=... |
apps/web |
Next.js dashboard — Lighthouse-style web UI |
Agent Readiness Score (0-100)
├── Discovery (25%) — llms.txt, robots.txt, sitemap
├── Identity (20%) — OG tags, JSON-LD, meta, title
├── Auth & Access (20%) — OAuth, OIDC, HTTPS, API docs
├── Integration (20%) — MCP, Webhooks, RSS, CORS
└── UX (15%) — Semantic HTML, ARIA, links, viewport
| Score | Grade |
|---|---|
| 80-100 | A — Fully Agent-Ready |
| 65-79 | B — Mostly Agent-Ready |
| 50-64 | C — Partially Agent-Ready |
| 35-49 | D — Minimally Agent-Ready |
| 0-34 | F — Not Agent-Ready |
- Monorepo: Turborepo + pnpm workspaces
- Core: TypeScript, native fetch
- API: Hono
- Web: Next.js 15, React 19, Tailwind CSS v4
- CLI: Commander, Chalk
MIT