Instantly search files, documents, code, and screenshots locally with OCR and MCP agent access.
Search filenames, full text, meaning, and screenshots instantly and privately.
Kensaku is a fast local file launcher and search tool written in Rust that indexes documents, code, and images with OCR. It suits power users, developers, and executive assistants who need to recall files by content, meaning, or text trapped in screenshots without sending data to the cloud.
It reads inside PDFs, Office documents, EPUBs, and over 50 code formats while providing semantic meaning search and natural-language time filters. It also exposes its local index to AI agents like Claude or Cursor via the Model Context Protocol (MCP).
It suits professionals managing thousands of local documents who want lightning-fast recall without privacy compromises. It does not suit users looking for a purely cloud-synced search engine or those unwilling to install standalone desktop binaries.
Extracts searchable text from screenshots and images using bundled or selectable OCR engines.
Exposes your local file index directly to AI coding assistants and agents securely over MCP.
Built-in custom unigram and bigram tokenization handles Chinese, Japanese, and Korean text.
Find files using conceptual queries like 'the PDF about indexing' or natural time filters.
The tool is free under its MIT license, with no paid hosted options mentioned in the README. The only cost is running it on your own hardware.
879 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
These reviewed AI tools do the same job instantly — from our scored directory.
No. All file indexing, searching, and OCR processing happen entirely locally on your machine, and your index stays in your OS config directory.
You can point an MCP client to the CLI using the 'kensaku-cli mcp' command or run 'kensaku-cli serve' for HTTP connections.
Yes. Kensaku reads pixels in images and screenshots using OCR so you can search for error messages or text contained within them.
See Zekai first in Google