GeoSolver
Find where a photo was taken, with the reasoning shown
About GeoSolver
GeoSolver determines where a photograph was taken by reading the same evidence a skilled human geolocator would: road markings and which side traffic drives on, language and typography on signage, vegetation and climate indicators, architectural style, and how those combine against map evidence. What separates it from a black-box guess is that the reasoning is shown. For its primary audience, GeoGuessr players trying to improve, the explanation is the entire point, because a correct answer without reasoning teaches nothing.
The product is positioned first as a GeoGuessr coach, which is the honest framing of who uses it, with reported figures of 99.2% accuracy across more than 10,000 images analysed and 50 or more countries covered. Clue-based reasoning is the differentiator: rather than returning coordinates, it walks through which features pointed where. History is retained so past analyses can be revisited, and an MCP integration means the capability can be wired into an AI agent rather than only used through the interface. Pricing is a straightforward monthly or yearly subscription with a large annual discount and a seven day refund before paid use. The same team publishes other apps including Cardify and a voice tool.
Upload a photograph, a screenshot or a GeoGuessr round. GeoSolver analyses the visible evidence, working through roads and driving side, signage language and script, vegetation and terrain, architectural style and any other distinguishing features, then reasons toward a location against map evidence. The output includes both the location estimate and the clue-based reasoning that produced it, so you can follow the logic and internalise it for next time. History keeps previous analyses available for review. For developers, the MCP integration exposes the capability to AI agents. A seven day refund applies before paid use begins.
- •Clue-Based Reasoning Shown - Explains which roads, signs, vegetation and architecture pointed where, rather than returning a bare answer
- •GeoGuessr Round Support - Works directly from game screenshots, which is its primary use case
- •Reported 99.2% Accuracy - Across a stated 10,000+ images and 50+ countries
- •Analysis History - Past results retained and revisitable for study
- •MCP Integration - Available to AI agents as a tool, not only through the web interface
- •Multi-Signal Analysis - Roads, signage, vegetation, architecture and map evidence combined rather than one heuristic
- •Mobile App - Available beyond the browser
- •Seven Day Refund - Before paid use begins, with cancellation any time
For GeoGuessr players who want to get better rather than just get answers, which the reasoning output serves directly. Beyond gaming, photo geolocation has real uses in journalism, open-source research and verifying where an image actually came from, and the MCP integration suggests developer use cases too. Travellers can identify where an old photograph was taken. One thing to be clear-eyed about: the same capability applied to a photograph of a person can reveal where they were, so it deserves the same care as any location data.
