GeoInfer is a tool that estimates where a photo was taken from a single image. It analyzes the visual content of the frame rather than relying on EXIF data or geotags, which is useful when files are anonymized, stripped of metadata, or cropped.
Geolocation from visual cues
The model looks for signals such as architecture, terrain type, road markings, vegetation, textures, and other scene details. Based on these cues, GeoInfer returns a likely location and can narrow the search to a specific region or city.
Built for OSINT and security workflows
GeoInfer is aimed at open-source intelligence practitioners, law enforcement, security analysts, and researchers. It can support common verification and analysis tasks, including:
- Checking photo authenticity and source credibility
- Matching reported events to real-world terrain
- Adding location context alongside other data sources
Global coverage and demos
GeoInfer supports global coverage, so you can upload images from anywhere in the world. The site also provides demo images to help you evaluate geolocation quality and see how the AI interprets different scene types. A public beta is available with a base model, with additional higher-accuracy options.
