Add a public API wiki at /wiki

Builds docs/ into a static site via mkdocs + mkdocs-material in its own Docker build stage (docs-builder), so mkdocs itself never ends up in the runtime image -- only the rendered HTML does, mounted at /wiki alongside the existing /docs OpenAPI schema (unconditionally, unlike /docs, since this is curated content meant to be public regardless of AURORA_API_DOCS_ENABLED). docs/aurora-core-spec.md (the internal build spec, referenced by "section N" comments throughout the codebase) is excluded from the built site entirely via exclude_docs, not just left out of the nav -- it was never meant for public consumption. Co-Authored-By: Claude Sonnet 5 <[email protected]>
2026-09-06 06:04:59 -05:00
parent 4403305da6
commit 56a03f9cfe
7 changed files with 113 additions and 1 deletions
@@ -0,0 +1,33 @@
+# Aurora Core
+
+Aurora Core is the metadata backend behind [AMDBS](https://www.amdbs.net) — a
+TMDb/IMDb-style canonical data layer for movies, TV, people, and everything
+that connects them.
+
+This wiki documents the public, read-only HTTP API. If you're building
+against Aurora Core, start with the [API Reference](api.md).
+
+## Base URL
+
+```
+https://api.amdbs.net/api/v1
+```
+
+All responses are JSON. Every entity has a stable, type-prefixed id (e.g.
+`aur_movie_01m1q7p1c...`) — always use the id a response gave you rather
+than constructing one yourself.
+
+## Interactive docs
+
+A live, explorable OpenAPI schema (Swagger UI) is available at
+[`/docs`](/docs) alongside this wiki, when enabled in the current
+environment.
+
+## Where to go next
+
+- **[API Reference](api.md)** — every endpoint, grouped by resource, with
+  query parameters and real example responses.
+- **[Architecture](architecture.md)** — how Aurora Core is put together:
+  ingestion pipeline, caching, provider abstraction.
+- **[Database](database.md)** — the schema itself, table by table.
+- **[Entity relationships](erd.md)** — the same schema as a diagram.