Building tools for geographic reasoning.
MapClue is a platform for AI-powered geographic reasoning. We believe that understanding the visual language of the world — its landscapes, architecture, flora, and culture — is a learnable skill, and that AI can accelerate that learning dramatically.
Our mission
The world is endlessly readable, but most people don't know where to look. A road marking, a plant species, a roof profile, a license plate shape — these aren't random details. They're geographic fingerprints. Each one narrows the possible answer.
MapClue builds tools that teach you to read those fingerprints. We use AI not to replace human geographic reasoning, but to make it visible, teachable, and transferable — so every session leaves you more capable than the last.
We started with GeoGuessr players, because they're the most motivated geographic learners in the world. But the applications extend to researchers, educators, travelers, and anyone who wants to understand the visual geography of the planet.
AI as a teacher, not a crutch
Every MapClue tool is designed to expose its reasoning — not just give answers. When you understand why the AI made a particular call, you build the intuition to make that call yourself.
Privacy as a first principle
We don't want your data. GeoSherlock is built so that your images go directly to the Anthropic API — MapClue never sees them. This isn't a policy choice; it's an architectural one.
Calibrated honesty
Geographic analysis is probabilistic. We don't hide that. When evidence is thin, the tool says so with honest confidence scores — rather than projecting false certainty.
Our approach
MapClue is built on the conviction that geographic reasoning is a structured skill — and that structured skills can be systematically taught. We don't just ask the AI "where is this?" — we ask it to reason through each category of visual evidence, weigh the signals, and explain its conclusions.
That reasoning process — vegetation → architecture → signage → roads → conclusion — is itself the teachable artifact. By making it visible, we let users internalize the method. Over time, players who use GeoSherlock in learning-oriented modes consistently report that they start noticing clues they previously overlooked, even without the tool open.
We use the most capable available vision AI for this reason. Claude Opus 4.6's ability to reason across multiple visual domains simultaneously — rather than classifying individual objects — is essential to the quality of analysis we aim for.
What we're building
GeoSherlock is the first MapClue product. More are in development.
GeoSherlock
AI-powered browser extension for geographic clue analysis from street-level photography. Available now on Chrome.
Region Atlas
An interactive reference for geographic visual signatures — searchable by country, clue type, and feature category.
Performance Tracker
Track your geographic accuracy across sessions, identify weak regions, and measure skill growth over time.
MapClue
MapClue is an independent software company building AI-powered tools for geographic reasoning. We're a small team with a focused mission: make the visual language of the world learnable.
Start exploring with GeoSherlock.
Create a free MapClue account and get access to GeoSherlock — and every product we release.