QuOTE questions in the enrichment pipeline + one-click eval default (#88)

Extractive synopsis: lead every api chunk with the page's own summary sentence

Contextual Retrieval: LLM glosses for api reference pages

Cut oversized sections at natural seams; embed only the new parts

Enrich indexed text with symbol + heading path (give URL/title semantic weight)

Switch embeddings to a local CPU model (bge-small): free tier APIs can't index the corpus

M2 indexing: Neon schema, resumable Gemini embedding, one-command build script