feat: add search

PROJECT_CONTEXT.md

@@ -17,6 +17,7 @@ Content is sourced from the pre-1980s Vedabase, extracted and indexed into Mongo
 - **vaul** for mobile drawer navigation
 - **Go BFF API** — all data fetched via `lib/bff.ts` HTTP client calling the Go BFF server (no direct database access)
 - **framer-motion** for audio player animations (spring enter/exit, minimize/expand transitions, audio wave bars)
+- **Typesense** (via BFF) — typo-tolerant full-text search across all content
 - **Python 3** with BeautifulSoup4 + `indic-transliteration` for HTML extraction scripts
 
 ---
@@ -69,9 +70,13 @@ knowledgebase/
 │   │   ├── place/[place]/page.tsx → Place listing
 │   │   ├── type/[type]/page.tsx → Type listing (lecture/conversation), year sidebar
 │   │   └── [slug]/page.tsx     → Individual transcription reading page
+│   ├── search/                 → Search page (server + client components)
+│   │   ├── page.tsx            → Server component (URL-driven state)
+│   │   └── search-content.tsx  → Client component (input, filters, results)
 │   └── api/
 │       ├── og/route.tsx        → Dynamic OG image generator (Noto Sans, Prabhupāda image)
 │       ├── books/              → Book/verse API routes
+│       ├── search/route.ts     → BFF search proxy (Typesense via /api/v1/search)
 │       └── glossary/route.ts   → Glossary API
 ├── components/
 │   ├── ui/                     → shadcn components (button, card, badge, tooltip, input, drawer)
@@ -86,6 +91,7 @@ knowledgebase/
 │   ├── page-header.tsx         → Shared header with ReadingSettings (max-w-5xl)
 │   ├── reading-settings.tsx    → Theme/font/size/justify accessibility panel
 │   ├── mobile-nav.tsx          → Left-slide drawer for mobile sidebar navigation
+│   ├── header-search.tsx       → Header search button with framer-motion expand animation
 │   ├── site-footer.tsx         → Global footer (About + Home links)
 │   ├── audio-player/           → Site-wide floating audio player
 │   │   ├── audio-context.tsx   → React context + provider, manages <audio> element
@@ -542,6 +548,23 @@ SSR and DS have parent chapters with sub-chapters stored as individual `prose_ch
 - Implemented in all 24+ reader pages via shared `MobileNav` component
 
 ### Search
+
+**Search** (`/search` page) — Typo-tolerant, diacritic-insensitive full-text search across all content.
+
+- **Powered by Typesense** via the BFF (`/api/v1/search` proxied through `/api/search`)
+- **9 indexed content types** combined into 8 user-facing filters: Books (verses + prose), Letters, Transcriptions, Glossary, Essays, Journals, Magazines, Other Documents
+- **Multi-select filters** — click multiple type tabs to combine (e.g., Books + Letters); "All" deselects everything
+- **URL state sync** — `/search?q=surrender&type=books,letters` is shareable and bookmarkable
+- **Header search button** — `HeaderSearchButton` component on every page header expands into a search input on click; press `/` from anywhere to trigger
+- **300ms debounce** on input
+- **Highlighted snippets** — matched text wrapped in `<mark>` tags
+- **Typo tolerance** — "Krishna", "Krsna", "Krishn" all find Kṛṣṇa results
+- **Diacritic insensitivity** — `_ascii` shadow fields indexed alongside originals (e.g., "Yashodanandana" matches "Yaśodanandana")
+- **Prefix matching** — partial words like "sarv dharman parit" match `sarva-dharmān parityajya`
+- **Clickable result cards** — entire card navigates to the source page
+- Files: `app/search/page.tsx`, `app/search/search-content.tsx`, `app/api/search/route.ts`, `components/header-search.tsx`
+
+### Per-Page Filters
 - Books page: diacritic-aware search (typing "Krsna" matches "Kṛṣṇa")
 - Glossary page: full-text search across terms and definitions
 - Uses Unicode NFD normalization + manual IAST character stripping
@@ -620,6 +643,8 @@ ALLOW_RTF=true                         # Enable RTF viewer (development only)
 NEXT_PUBLIC_SITE_URL=https://prabhupadasays.org  # Public site URL for OG images
 ```
 
+> No new vars are needed for the frontend search feature — search goes through the BFF, which has `TYPESENSE_URL` and `TYPESENSE_API_KEY` configured server-side.
+
 ---
 
 ## How to Run
@@ -776,4 +801,6 @@ Content block formats: `"paragraph"` (body text), `"verse"` (centered italic), `
 
 7. **Empty transcriptions** — 30 entries have no English text content (Hindi/Bengali recordings); the UI shows "No transcription available" with the audio player still accessible
 
-8. **Audio player z-index** — Player is hidden when text selection toolbar is active on mobile to avoid overlap
+8. **Audio player z-index** — Player is hidden when text selection toolbar is active on mobile to avoid overlap
+
+9. **Search depends on Typesense** — Search depends on Typesense being available via the BFF; if Typesense is down, search will return empty results but the rest of the site is unaffected.