Haddon Institute for developers and agentsHaddon Institute for developers and agents

OpenAPI, API catalog, rate limits, and CLI quickstart for integrating with the public Haddon Institute website. Public endpoints are rate limited; use the headers below to self-throttle. Authenticated student systems (Populi) and staff admin APIs are out of scope here.

Quickstart

1. Read llms.txt for when to use this site.
2. Fetch /.well-known/api-catalog or openapi.json.
3. Probe /api/health, then call documented routes (e.g. GET /api/search?q=admissions).
4. Optional CLI: npx haddon-institute health (see repository cli/haddon-institute.mjs).

Rate limits

POST APIs return X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset (Unix ms), and Retry-After (seconds) on HTTP 429. Chat allows 20 requests per 10 minutes per IP; form endpoints typically allow 3 per 10 minutes unless noted in OpenAPI.

Resources

• OpenAPI 3.1 — Machine-readable HTTP surface (operationId on each operation).
• API catalog (RFC 9727) — Linkset discovery for search, chat, auth, and health.
• HTTP API overview — Human-readable endpoint table and scope notes.
• llms.txt — Site map hints and when-to-use guidance for agents.
• Agent instructions — When to call Haddon vs other sources.
• ARD ai-catalog.json — Capability manifest (MCP card, OpenAPI, chat, catalog).
• auth.md — Agent authentication and registration policy.
• OAuth protected resource (RFC 9728) — How to obtain tokens for protected /api routes.
• OAuth authorization server (RFC 8414) — Issuer, endpoints, and agent_auth block.
• MCP server card — Transport endpoint and capabilities for site chat.
• Health check — JSON `{ "status": "ok" }` for uptime probes.

Markdown pages

Send Accept: text/markdown with a higher quality value than text/html on any public HTML page (including the homepage) to receive a Markdown representation with Vary: Accept.