kensaku
Search inside documents, code, and even screenshots - instantly and locally. A free, open-source file launcher with OCR, semantic search, and MCP agent access.
What is kensaku?
Search inside documents, code, and even screenshots - instantly and locally. A free, open-source file launcher with OCR, semantic search, and MCP agent access.
Overview of kensaku
What it is
Kensaku is a fast, lightweight, and efficient file launcher that searches inside documents, images, and code locally. It is free and open-source under the MIT license. It can search 100,000 file names in about a millisecond using roughly 12 MB of memory. Nothing leaves your machine.
Key features
- Filename search across ~100,000 files in a few milliseconds on a warm cache.
- Full text search inside PDF, DOCX, PPTX, XLSX, EPUB, ODT/ODS/ODP, RTF, plain text, code, and logs, including CJK text.
- Text inside screenshots and images via OCR (PP-OCRv6 by default, optional Indra engine and Tesseract).
- Semantic search: find by meaning (e.g., “the PDF about indexing”).
- Natural-language time queries (e.g., “screenshots from last week”).
- MCP (Model Context Protocol) support for AI agents to search the index read-only.
- Launcher behavior: starts hidden in tray, summon with keystroke, results update as you type.
Use cases
- Searching for files when you remember content but not the filename (e.g., a PDF about database indexing saved as notes-final-2.pdf).
- Finding log lines from failed deployments in a folder of logs.
- Searching screenshots for error text (e.g., a CUDA error).
- Describing files with natural language (e.g., “big videos from last month”).
- Giving AI agents (like Claude or Cursor) local search capabilities via MCP.
Getting started
bash git clone https://github.com/yjid/kensaku cd kensaku bun install bun run tauri build
License and pricing
MIT license. Open source and free.
Based on the project's public README and repository metadata.