OSINT Reverse Image Search
Reverse Image Search lets you start from a single photo —of a person, a building, a product or a place— and find out who or what it shows, where it was taken and which web pages it appears on. It combines Google Cloud Vision web-entity recognition (which links the image to known people, brands and places), landmark detection with geographic coordinates, face and logo analysis, and extraction of the GPS embedded in the photo's own EXIF metadata. It is one of the most powerful OSINT techniques for geolocating imagery and identifying subjects.
What it does
- Identifies people, brands and places via Google Vision web entities and "best guess".
- Finds every web page where the same image appears (where it has been published).
- Landmark and place detection with latitude/longitude and a direct Google Maps link.
- EXIF GPS extraction (the photo's real location) with no API key required.
- Face, logo and visually-similar image detection to pivot the investigation.
How to use it
- 1Upload an image or paste it directly from your clipboard.
- 2Run the search: it is analysed with Google Vision and the EXIF is parsed locally.
- 3Review the likely identity, the location on the map and the sites where the image appears.
Use cases
- People investigations: identify someone from a profile photo.
- Geolocation (geoint) of photographs and verifying where they were taken.
- Content verification and detecting reused or fake images (fact-checking).
Frequently asked questions
Do I need an API key?
Full identification uses Google Cloud Vision (a free key with 1,000 uses/month that you set in Settings). EXIF GPS extraction works with no key at all.
Can it identify anyone?
It recognises people and subjects with public, web-indexed presence (celebrities, well-known profiles). For anonymous people it returns matching pages and similar images so you can keep investigating.
How does it find the location?
It combines Google Vision landmark detection (with coordinates) and any GPS the camera embedded in the photo's EXIF metadata.