corpus-search

What problem does it solve?

You need reliable ways to find relevant information inside large collections of markdown notes, transcripts, and summaries without manually reading everything.

Core Features & Use Cases

• Hybrid retrieval over your corpus: combines keyword matching and semantic search to return ranked, relevant chunks.
• Local, sandbox-only index: builds a persistent SQLite store (including FTS for BM25 and vector search for embeddings) with no separate server to run.
• Useful metadata for real workflows: supports role filtering and timestamp extraction from markdown so results can be traced back to where they occurred.

Quick Start

Ask to install and index your markdown sources, then immediately run a natural-language query against the indexed kind to get ranked results back in under a second.