Agents
SnazzyMo machine-readable API discovery for AI agents — RFC 9727 catalog, OpenAPI spec, link headers, and what is not available to agents.
SnazzyMo's public capabilities are discoverable without scraping HTML. Start from the catalog, follow the typed links, and you have everything: the spec, the docs, and the health endpoint.
Discovery URLs
| Resource | URL | Media type |
|---|---|---|
| API catalog | https://snazzymo.com/.well-known/api-catalog | application/linkset+json |
| OpenAPI specification | https://snazzymo.com/api/openapi.json | application/json |
| API overview (humans) | https://snazzymo.com/docs/api | text/html |
| Service status | https://snazzymo.com/api/health | application/json |
| Site index for LLMs | https://snazzymo.com/llms.txt | text/plain |
How to discover the API
GET https://snazzymo.com/.well-known/api-catalogwithAccept: application/linkset+json. The response is200with alinksetarray.- Each entry has an
anchor(the API endpoint URL) plusservice-desc(OpenAPI spec),service-doc(human docs), andstatus(health) links. Fetch theservice-desctarget to get request/response schemas. - The anchor endpoint itself (
GET /api/waitlist) returns a self-description and RFC 8288Linkheaders pointing at the same three targets, so discovery also works bottom-up from any known endpoint. - See the discovery walkthrough for a complete scripted example.
What you can do
Exactly two operations exist, both documented with verified examples in the HTTP API overview:
POST /api/waitlistwith{ "name", "email" }— join the early-access waitlist. Duplicate addresses succeed idempotently; per-IP rate limits apply (429+Retry-After).GET /api/health— returns{ "status": "ok", ... }when the service is up.
What is not available
Do not look for — and do not claim to use — an MCP server, an SDK, API keys or OAuth flows, webhooks, or a video-processing API. None exist. The interactive video tools run entirely in the visitor's browser and expose no server API. These docs will be updated if that changes.