illegal-instruction-co

illegal-instruction-co / recall-lite

Public

semantic search for your local files find by meaning, not keywords. 120+ file types, OCR, MCP server for AI agents. 100% private.

ai-agents embeddings file-search knowledge-management lancedb
53
3
100% credibility
Found Feb 17, 2026 at 20 stars 3x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Rust
AI Summary

A desktop app for Windows that enables semantic search across local files, supporting text extraction from PDFs and images via OCR, EXIF metadata parsing, and hybrid search methods, all running offline.

How It Works

1
🔍 Discover smart file search

You're frustrated with slow file searches and find Recall Lite, a helpful tool that understands what you mean when looking for files.

2
💻 Get the app running

Download the app and start it up – it quietly sits in your system tray, ready whenever you need it.

3
⌨️ Pop open the window

Press Alt + Space anywhere on your computer to bring up the search window instantly.

4
📁 Pick your folders

Choose the folders with your documents, photos, code, and more so it knows where to look.

5
⏳ Let it scan and learn

The app quickly goes through your files, understanding their content, even reading text in images and photo details.

6
🔤 Search in your own words

Type natural phrases like 'summer photos from Istanbul' or 'that Python function' and see matching files pop up with previews.

🎉 Find and open instantly

Click or press Enter to open the perfect file right away, making your computer feel organized and fast.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 20 to 53 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is recall-lite?

Recall-lite is a desktop app built with Rust and Tauri that delivers local semantic search for your Windows files, fixing the limitations of built-in Windows search. Pop it in your system tray, hit Alt+Space to toggle, type natural queries like "summer morning photos from Istanbul," and it surfaces files via hybrid vector, full-text, and reranker matching. It indexes 50+ formats including code, PDFs, images (with OCR and EXIF parsing), and supports isolated "containers" for work/personal splits—all offline, no cloud phoning home.

Why is it gaining traction?

Unlike cloud-dependent tools like Copilot, recall-lite runs fully local with embedded LanceDB and multilingual embedding models, pulling in reverse-geocoded locations and human-readable dates from EXIF for intuitive photo searches. Devs dig the smart chunking for code (splits at functions/classes) and docs (headers/keys), plus query expansion that strips stop words in English/Turkish. Global shortcuts, incremental indexing, and i18n make it feel native on Windows, beating Spotlight's literal keyword hunts.

Who should use this?

Windows power users drowning in unsearchable file hoards, especially devs indexing Rust/Python/JS codebases or Markdown notes alongside configs/logs. Photo collectors benefit from "halo bassinest lite recall"-style semantic EXIF queries, while remote workers need container isolation for client projects without remnants. Ideal if you're tweaking windows github actions or ssh keys and hate digging through folders.

Verdict

Try recall-lite if Windows search aktivieren fails you—it's a solid local alternative with MIT license and easy npm run tauri dev setup. At 11 stars and 1.0% credibility, it's early-stage with thin docs and no broad tests, so expect tweaks; still, promising for offline recall literacy over jeep 4 liter recall hunts.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.