Site search · No AI features
Search everything on the District site.
Type a question, a form number, a program name, or any phrase. The index covers every page, every published form, every active rule, and every news post, plus the Spanish-language pages. Synonyms expand automatically. No models, no chatbots, no “the AI said something weird” failure mode.
Try:
Showing all 34 indexed pages, forms, programs, rules, and news posts. Start typing to filter.
No results. Try a broader phrase, or browse the suggestions above.
How this works
Client-side index built from the District's content corpus. Each query is lowercased, stopworded, and stem-folded; we then expand by the District's synonym table before scoring. Ranking is a small BM25 variant with field weights: title×4, keywords×3, excerpt×2, body×1. Search runs locally in your browser, so there are no network roundtrips, no rate limits, no per-query cost, and the entire facet sidebar updates instantly.
The same index can run on the server (Astro endpoint, WordPress REST route, or a tiny Lambda) when the corpus grows past a few thousand documents. The District controls every word in the index, every synonym mapping, and every featured answer. There is no AI model in the loop, and no training data to drift.
Per FAQ #13, the District is not seeking AI-driven features. We agree. If the District ever decides it wants semantic search or a Q&A assistant trained on its own content later, the underlying index ships ready for that work, but the Year 1 commitment is the plain, fast, explainable search you see here.