Research, field notes, and visual intelligence.
How Graylark builds, evaluates, and deploys visual intelligence for consequential investigative work.
Why We Built GeoSpy and Why It Became Raven
GeoSpy started as a two-week experiment in a one-bedroom Boston apartment. The investigators who found it showed us what it needed to become.

Where Was This Photo Taken? How AI Detects Locations From Visual Clues
Wondering where a photo was taken? Learn how modern AI photo location finders analyze architecture, roads, foliage, and lighting to pinpoint locations without metadata.
Read article
How to Find Photo Location Without Metadata (EXIF Data)
Learn how to locate where an image was taken even when EXIF GPS metadata is completely stripped. Step-by-step investigative methods using AI geolocation, visual clues, and reverse tools.
Read article
Top 5 Visual Clues That Reveal Location in Any Photo
Discover the top 5 visual indicators experts and AI models use to geolocate photos: architecture, road markings, flora, utility infrastructure, and sun shadows.
Read article
AI Geolocation vs. Reverse Image Search: Which Works Better?
Reverse image search vs AI photo geolocation: learn the fundamental differences, why Google Lens fails on personal photos, and how AI finds locations from new images.
Read article
The Science Behind AI Photo Geolocation: Neural Networks & Geographic Embeddings
Explore the deep learning architectures, Vision Transformers, and hierarchical geocells that enable AI systems to estimate photo locations across Earth in milliseconds.
Read article
Geolocating a Suspected Stolen Car From a Single Social Media Post
One public listing, one low-context photo, no metadata. How Raven takes an investigator from a suspicious vehicle post to a street-level hypothesis, a visual match, an organized case, and nearby camera leads.
Read article
A Research Lab in Service of Its Customers
Graylark combines the rigor of defense research, the speed of Silicon Valley prototyping, and product design that makes advanced AI usable in the field.
Read article
What Is Image Geolocation? How Investigators Locate Photos Without Metadata
What image geolocation is, how it differs from GPS and reverse image search, where it breaks down, and how investigators turn a low-context image into a verified location.
Read article
Graylark Raises $10.7 Million to Build Raven
The seed round was led by General Catalyst, with backing from AI Grant and participation from Alexandr Wang, Recorded Future, and other investors and operators.
Read announcement
GeoSpy 101: AI Image Geolocation, Explained
What GeoSpy is, how visual geolocation works, and how it evolved into Raven by Graylark.
Read articleSee Raven run on your own imagery.
Raven is available to verified agencies and investigative teams.