One-fetch, site-wide discovery for AI agents — a single JSON file at a well-known endpoint.
AI-Me defines a JSON document served at /.well-known/ai-me.json that provides canonical pages, offerings, pricing model, and primary calls-to-action in a strict, validatable shape. It complements existing efforts like llms.txt, schema.org, MCP, and A2A.
Website: spec.ai-me.dev
- Create
/.well-known/ai-me.jsonon your site:
{
"schema_version": "0.1",
"name": "Your Company",
"website": "https://example.com",
"summary": "What you do, in one sentence.",
"canonical_pages": {
"home": "https://example.com/",
"contact": "https://example.com/contact"
},
"last_updated": "2026-03-05"
}-
(Recommended) Add
/.well-known/llms.txtwith a pointer to your vendor card. -
Validate:
npx @ai-me/validator https://yourdomain.comai-me/
├── spec/ Normative specification markdown
├── schemas/ JSON Schema for validation
├── examples/ Example vendor cards (SaaS, consulting, OSS, e-commerce)
├── packages/
│ └── validator/ CLI validator tool
├── website/ Astro-based documentation site
└── docs/ Supporting documentation
# Install dependencies
pnpm install
# Run the website locally
pnpm dev
# Build the website
pnpm build
# Preview the built site
pnpm previewUsing AI-Me in production? Open a PR to add yourself here.
Be the first to adopt the spec and get listed here.
This project uses a dual-license structure:
- Specification, schemas, and examples — CC-BY 4.0
- Code (website, validator, tooling) — Apache 2.0
See LICENSE for full details.
