Privacy Policy

Last updated: March 2026

The short version

PixaFind runs on your machine, and your files never leave your computer. During the public beta (before v1.0), the app sends crash and error reports so we can fix bugs quickly — this will be removed in the stable v1.0 release. Apart from that and optional update checks, PixaFind makes no network calls, and it never uploads your files, photos, or their contents.

Data collection

We don't collect personal data or sell anything. Your files, photos, and their contents are processed entirely on your device and are never uploaded. All AI models run locally on your hardware. The one beta exception is crash and error diagnostics (see below), which contain technical error data only.

Crash reporting (beta only)

To reach a stable v1.0, beta builds include Sentry crash & error reporting. It captures diagnostic information when something breaks — error messages and app state — so we can fix it. It does not include your files, photos, or their contents. This is temporary: the stable v1.0 release will ship with zero telemetry, matching the "no network calls" goal.

Your files

Files you index remain on your local machine. The search index (embeddings and metadata) is stored locally in your app data directory. No file contents, filenames, or metadata are ever transmitted anywhere.

AI models

All AI models are downloaded once on first launch and run locally. Model inference happens entirely on your CPU/GPU. No API calls are made to any cloud AI service.

Updates

The app checks GitHub releases for available updates. This is the only network activity, and it can be disabled. No personal or usage data is included in these requests.

Website

This website uses privacy-friendly Google Analytics to measure traffic. Analytics cookies are set only after you accept via the cookie banner — decline and no analytics cookies are stored. This applies to the website only; the PixaFind app itself collects nothing.

Contact

If you have questions about this privacy policy, reach us at our contact page or open an issue on GitHub.