What the tool looks at
- EXIF data. GPS position, local time, camera, phone model and editing software, when the file still has them.
- Signs and text. Latin, Cyrillic, Chinese, Japanese and Arabic by default, with Korean, Hindi, Greek and Thai on request. Text is matched against brands, phone codes, web addresses, currencies, postcodes and place names.
- Visual clues. Scenery, plants, building styles, vehicles and road markings, each linked to a guide or an image search.
- AI location guess. GeoCLIP, a 321 MB download that only runs after you agree GeoCLIP ONNX model card (read 21 Nov 2026).
- Street crossings and sun angle. Two street names give an exact corner. A shadow and a time narrow the latitude.
- Map lookups. Shop names, postcodes, phone area codes, distance signs and combinations of nearby features, as described on the Video Location Finder page.
- Consistency checks. GPS, clock offset, brands, light, snow, dates and the weather record are compared with each other, as explained here.
- One click, one report. Every ticked check runs in order and ends in a findings report, as described here.
Why add several photos?
Clues add up. One photo may show a shop name, another a road sign, a third the hills. The tool ranks places that fit all of them. Only add photos taken in one place.
How it fits with the Image Verification Workbench
The Image Verification Workbench checks whether a photo is real and where it first appeared. This tool asks where it was taken. Run both on an important photo.
What leaves your browser
Your photos never leave the tab. Model files come from Hugging Face and jsDelivr at fixed versions and are checked against SHA-256 hashes. With “Online checks” ticked, short sign text, street names and postcodes go to Photon, Overpass and Zippopotam.us. One date and position go to Open-Meteo, and Wikimedia Commons is asked for photos near the top lead. The picture itself is never sent. Untick “Online checks” to keep everything in the browser; the buttons still run each check on its own.
Limits
Text readers misread small or angled signs. Image labels are hints, not facts. GeoCLIP is strongest on well-photographed places GeoCLIP paper, arXiv 2309.16020 (read 21 Nov 2026). EXIF can be edited, and social networks strip it.
Models and data
| Part | What it does | Download | Licence |
|---|---|---|---|
| PP-OCRv5 mobile | Finds and reads text; one reader per script | 13 MB, plus 8 to 17 MB per extra script | Apache 2.0 |
| SigLIP base | Visual clues: scenery, plants, buildings, vehicles, roads, script style | 100 MB | Apache 2.0 |
| GeoCLIP | AI location guess over 26,958 map cells | 307 MB model + 14 MB index | MIT |
| Whisper tiny | Spoken language | 41 MB | Apache 2.0 |
| GeoNames | 34,248 towns and cities with 15,000 or more people, 2,548 region names, and country facts | 2.7 MB, with the page | CC BY 4.0 |
| Natural Earth | Country outlines for the map | 0.1 MB, with the page | Public domain |
| libphonenumber geocoding data | 14,079 phone area codes in 45 countries | 0.5 MB, with the page | Apache 2.0 |
| Name Suggestion Index | 43,683 brand, operator and transit names and the countries they trade in | 2.7 MB (0.6 MB compressed), with the page | BSD 3-Clause |
| IANA time zones | Which countries used a clock offset on a date | With the page | Public domain |
| OpenStreetMap via Photon and Overpass | Street crossings, sign names and map searches, with “Online checks” ticked, or when you press a button | None | ODbL |
| Zippopotam.us | Postcode and town checks, with “Online checks” ticked, or when you press a button | None | Service |
| Wikimedia Commons | Geotagged photos near the top lead, with “Online checks” ticked, or when you press a button | None | Per photo |
| Open-Meteo historical weather | Rain, snow and cloud on the day, with “Online checks” ticked, or when you press a button | None | CC BY 4.0 |
Frequently asked questions
Does it work on screenshots?
Yes, if the screenshot shows the scene. Screenshots carry no camera GPS, so the tool relies on text and visual clues.
Are my photos uploaded?
No. Everything runs in your browser tab. With online checks on, short sign text, street names, a date and map positions go to map and weather services, never the photos.
Why is the AI location guess off by default?
It is a 321 MB download and needs about 1.5 GB of memory. The tool asks before it downloads it.
Can it identify the people in a photo?
No. This tool only looks for place clues. It never matches faces.