Docs: API & developer resources
mattwood.fyi is a typed knowledge graph of riffs, links, and essays — designed to be read by people and queried by agents. Everything here is public: no authentication, no API keys, no rate limits. All API responses are JSON.
Machine-readable entry points
- /openapi.json — OpenAPI 3.1 description of the query API (operationIds, typed parameters, response schemas).
- /llms.txt — llms.txt index of the whole site.
- /feed.json — JSON Feed 1.1 with a
_fyiextension per item (recommended single source for agents). - /feed.xml — Atom feed with full HTML content.
- /sitemap.xml — every indexable URL.
- /agents/ — copy-paste agent instructions and when-to-use guidance.
Query API
Path-based endpoints are preferred because some agent fetch tools strip query-string parameters that look like IDs.
GET /api/fyi/q/search/{keyword}— keyword search across title and content.GET /api/fyi/q/semantic/{query}— semantic (vector) search by meaning; results ranked by similarity.GET /api/fyi/q/items?since=YYYY-MM-DD&type=link|riff|essay&limit=N&offset=N— list/filter items with pagination metadata.GET /api/fyi/q/edges/{itemId}— all connections for one item (type, direction, confidence, reason). Add?type=to filter.GET /api/fyi/q/edges?type=challenges— all edges of one type across the graph.GET /api/fyi/q/summary— pre-computed snapshot: top connected items, recent activity, edge distribution, tensions.
Example
curl https://mattwood.fyi/api/fyi/q/semantic/agent+harness+architecture
For the full field reference and function-calling-ready schemas, use the OpenAPI spec. For narrative guidance on when to reach for this site, see the agent instructions.