வெண்முரசு

• Search — Deep Research compatibility: hybrid search over the Venmurasu novel (Tamil or English query). Returns passage, entity, and chapter results with stable ids for the `fetch` tool. MCP clients that want richer response shapes should prefer `search_passages` instead. PERFORMANCE: supply `book` when you know the book of interest and `max_chapter` when the reader is mid-book - both narrow the search to that subset and cut response time from ~7s (full corpus) to ~1s. If the user mentioned a specific boo
• Fetch — Deep Research compatibility: fetch full content for an id returned by `search`. Accepted id shapes: `passage:{book}/{chapter}/{p_id}`, `entity:{id}`, `chapter:{book}/{n}`, `book:{slug}`. Returns the canonical user-facing URL so citations link back to the reader.
• List Books — List all books in the Venmurasu epic series (26 books, ~26,000 pages of Tamil literary fiction by Jeyamohan). Returns title, slug, and chapter count for each book.
• Get Book — Get details for a specific Venmurasu book including its full chapter list with section names.
• Read Chapter — Read the full text of a chapter as markdown. The content is Tamil literary fiction.
• Lookup Word — Look up a Tamil word in the Venmurasu dictionary. Returns the definition if found.
• Search Passages — Hybrid semantic + keyword search over Venmurasu passages across all ingested books. Combines BM25, Tamil halfvec embeddings, and English-bridge embeddings in one query. Input can be Tamil or English. Use this for narrative queries like 'Bhishma's vow', 'snake sacrifice', or 'creation of the universe'. Returns ranked passages with chapter anchors.
• Get Entity — Look up a single Venmurasu entity (character, place, event, theme, etc.) by canonical id. Returns canonical name, kind, bio, first appearance, aliases, and mention count.
• List Entities — Browse Venmurasu entities filtered by kind and/or first-appearance book. Useful for directory views or finding all characters/places/themes in a book.
• Get Entity Context — Fetch rich narrative context for a single entity: the entity bio, chapter arc (mention density per book+chapter), and the top-5 'key scenes' ranked by vector similarity of each passage to the entity's canonical name. Use this to answer 'tell me about X' / 'what role does X play' questions in one call.
• Explain Query — Fast 'what did the search backend understand' preview for a query. Returns only the interpretation block (translated English, resolved entity count, channels fired, confidence). No passages, no entity details. Use it to cheaply validate a Tamil/English query before committing to the heavier search_passages call.
• Persona Distribution — Operator tool (v1.0.0). Current persona split (reader/social/scholar) from Supabase user_persona_state. Bearer-auth required. Args: window_days, region (ISO country), include_holdout, format.
• Persona Switch Rate — Operator tool (v1.0.0). % of users who switched personas in the last N days. Reads persona_history. Bearer-auth required.
• Persona Retention — Operator tool (v1.0.0). Reader completion rate per persona. Joins user_persona_state + venmurasu_reader_sessions. Bearer-auth required.
• Persona Suggestion Coverage — Operator tool (v1.0.0). Of users in session_summary, how many have a suggestion computed within 48h. Bearer-auth required.
• Persona Invite Activity — Operator tool (v1.0.0). Top invite codes by redemption count, geo spread, persona mix. Bearer-auth required.
• Persona Funnel — Operator tool (v1.0.0). Funnel: first visit → persona chosen → first content action. Bearer-auth required.
• Persona Explain — Operator tool (v1.0.0). Debug tool: given a device_id, returns the computed suggestion + top weighted signals + reason each signal mapped to that persona. Bearer-auth required.
• Search Dictionary — Search the Venmurasu Tamil dictionary. Matches words or definitions containing the query text.

Schemas and invoke metadata: Directory API