your photo, not ours
Geolocation tools sit in an uncomfortable place, so we made the uncomfortable thing the design: the photo is processed, reasoned about, and forgotten. It is never stored.
EXIF stays on your device
GPS, camera and timestamp are parsed in JavaScript inside your browser. The raw file never leaves your machine for that step.
the vision pass is a thumbnail
If there's no GPS, we send a downscaled copy (max 768px, JPEG) to a vision model on our edge — not the original, and not to any third-party analytics.
we store coordinates, not images
The only thing we keep — and only when you run a solve — is a coordinate, a method, a confidence number, and a timestamp. No pixels, no faces.
no accounts, no tracking
There's no login and no cross-site advertising ID. We don't know who solved what; we only know that a photo was solved.
what "nothing stored" actually means
When you drop a photo, this is the full path it takes:
- the bytes are read into memory in your browser;
- the EXIF is parsed there, locally;
- if the model is needed, a downscaled thumbnail is sent to our edge function and immediately discarded after inference;
- if a location is found, only the coordinate, method and confidence are written to our database;
- the page keeps a rendered preview in your own browser's memory, nowhere else.
At no point does the original photograph exist on a server we control.
an ethical note
sleuth is for understanding your own photos — where a memory happened, or verifying the provenance of an image you received. It is not for locating or tracking people without their knowledge. Don't use it for that. If you do, you're on your own — and we'd rather you didn't use it at all.