API Documentation

Live view of the backend's /openapi.json spec — auto-updates as endpoints change on the server. Tags group endpoints by router (analytics, pages, similarity, etc.).

⏳ Loading spec…

Consuming from external tools: the backend is reached via the n8n proxy (POST {path, method, body} to the proxy URL). Stable endpoints for external use: /sites, /pages, /similarity/pairs, /similarity/clusters, /analytics/other-urls, /export/*.

Sprint 34 — cross-tool integration endpoints (stable, documented in INTEGRATION-NOTES-*.md files in consuming tool repos):GET /sites/{site_id}/link-graph (paginated nodes + edges for site graph visualization with edge_kind filter + orphan inclusion),GET /sites/{site_id}/similarity-pairs (filtered pair rows for merge-overlay rendering with outcome + min_reranker filters),POST /pages/embeddings/search (semantic KNN over pages by query text; accepts optional persona_context), and POST /ideas/check (now accepts optional persona_context field prepended to the idea before embedding).

Internal-only: job-mutating routes (/crawl/*, /similarity/analyze, /analytics/gsc/sync, etc.) — these are safe to call but have non-idempotent side effects.