A phone clip often shows more than it seems: a shop sign, a street name, a skyline. Upload the recording and AI checks frames from it for clues to where it was filmed. When the clues aren't there, it tells you so instead of guessing.
Choose the clip and pay once. We pull 12 frames from it in your browser, have Google Gemini and Anthropic Claude check each one, and go with the place most frames agree on. The scan starts as soon as you return from checkout.
One payment, one scan, secure checkout by Stripe. No account or subscription. Not sure the clip has enough clues? Run the free 4-frame scan below first.
Phones can save a location inside the video file itself. An iPhone does it when Location Services is on for the Camera app; most Android camera apps have a similar location setting. If the original file still carries that data, it is more exact than anything read from the picture.
On an iPhone, open the video in Photos and swipe up: a map means the location was saved. In Google Photos, open the video and look at its details. On a computer, the free command-line program ExifTool shows it (exiftool -ee -G1 -a "-GPS*" video.mov).
Most copies have lost it. WhatsApp, Instagram, TikTok, Messenger and most other apps re-encode videos and drop the location when you send or post them, and screen recordings never had it. Our tool does not read video metadata; it works only from what is visible in the frames. That is the part that still works on a forwarded or downloaded clip.
With no location data, the answer has to come from the picture. These are the clues that actually decide it, roughly from strongest to weakest:
A phone video has one advantage over a photo: the camera moves. A pan or a short walk can catch a sign that no single frame of a photo would show, so sampling several frames gives more chances to find one clear clue.
| Free scan | Deep video scan | |
|---|---|---|
| Frames checked | 4, evenly spaced through the clip | 12, evenly spaced through the clip |
| AI models per frame | Our free model | Google Gemini and Anthropic Claude |
| How the answer is chosen | Place most frames agree on | Place most frames agree on |
| Price | Free, no sign-up | $2.99 for one video, one payment |
| Your video | Stays in your browser; only frames are sent | Stays in your browser; only frames are sent |
A deep scan gives a clip with some visible clues more chances and stronger models. It cannot add clues that aren't in the footage: if the clip is indoors, a close-up, or shot at night with nothing readable, a deep scan is unlikely to find more than the free one.
It doesn't track phones, read GPS from a video file, or find where someone is now. It only looks at the frames you give it. Results are usually at city or landmark level when a sign or landmark is visible, not a street address, and a clip with no distinctive features will come back with no confident location. Only use it on recordings you have a right to examine.
Often, if the original file still has its saved location (check it in your phone's gallery) or if the footage shows something identifiable, such as a sign, a landmark or a distinctive street. Upload the clip here and the AI checks frames from it; if nothing in the frames is distinctive enough, it says it found no confident location.
The original recording can, if location access was on for the camera. Copies sent through messaging apps or downloaded from social media almost always lose it. This tool does not read that metadata; check it in Photos, Google Photos or ExifTool.
It depends on what the clip shows. A readable street name or a known landmark can lead to the exact spot; architecture or landscape alone usually narrows it to a region at best. Always confirm the answer on a map.
One scan of one video: 12 frames instead of 4, each checked by Google Gemini and Anthropic Claude, with the answer chosen by agreement between frames. One payment through Stripe, no account and no subscription.
No. Frames are extracted in your browser and only those still images are sent for analysis. For a deep scan they are sent to the two AI models through our provider; see the privacy policy for details.