Recall — local multimodal semantic search for your files

What is Recall?

A new GitHub project called Recall bills itself as a local, multimodal semantic search for your files. The repository’s README says it indexes documents and other file types so you can search by meaning, not just filenames or keywords. It has been reported that the project was shared on Hacker News, where privacy-minded users and productivity tinkerers took notice. Want a private Google for your desktop? This pitches itself as something close.

Why it matters

The pitch is simple and appealing: keep your data local, query across PDFs, notes, images and more, and get results driven by semantic embeddings rather than brittle text matches. For people tired of sending their most sensitive documents to cloud services, that promise lands like a breath of fresh air. There’s a real emotional tug here — control, safety, and speed — especially as trust in big tech wobbles and local-first tools trend upward.

Caveats and context

The README outlines capabilities, but it has been reported that users should test and vet the project before trusting it with critical data — open-source doesn't mean turn-key or infallible. If you’re curious, the code and details live on GitHub (https://github.com/aayu22809/Recall), and community feedback on Hacker News can help you separate polished features from work-in-progress. Tinkerers will love it; everyone else should proceed with the usual caution.

Sources: github.com/aayu22809, Hacker News