kensaku
by yjid
Kensaku is a free desktop app that searches file names, the text inside documents and code, and text inside screenshots, all on your own computer. It works as a quick-summon launcher and can also let AI assistants search the same local index.
Hidden gemA free, local, cross-platform desktop launcher that searches file names, text inside documents, OCR'd screenshots and meaning, with a CLI and read-only MCP access for AI agents. It ships signed-off-by-checksum installers for three platforms, has documented benchmarks and a real query language, and goes well beyond filename-only tools like Everything. Best for anyone who loses files and wants local full-text and screenshot search.
New and promisingUnder two months old. Expect rough edges and fast changes.
- Stars
- 245 +44 a day, est.
- Licence
- MIT
- Runs on
- Your computer
- Setup
- Easy
- Language
- Rust
- Age
- 5 days
What you can do with it
- Find a PDF or Word document by remembering what was in it rather than its file name
- Search for error messages or text inside screenshots using built-in OCR
- Let Claude or Cursor search your local files through MCP, read-only
Could replace
- Alfred
- Listary
Our suggestion based on what it does, not a feature-by-feature comparison.
Ways a business could use it
- Quickly locate old invoices, contracts and client notes by their contents, saving time spent digging through folders
- Search scanned receipts and screenshots for figures or reference numbers when preparing accounts or answering client queries
Can you use it commercially?
Free to use, change and sell, including in paid products. Keep the copyright notice.
Not legal advice. Read the project's licence file before building a product on it.
Getting it running
Download the installer for Windows, macOS or Linux from the latest release page, then choose which folders to index on first run.
Follow the official instructions in the project README.
Things to check
- Builds are not code-signed, so Windows and macOS show security warnings on first run
- The index holds file names and extracted text, so it should be treated as sensitive
- Speed figures come from Linux only and comparisons with Everything and Listary are not measured
Checked 30 Oct 2026. Summary drafted by AI from the project's README.
