Raven demos locate homes from interior photos
Tech

Raven demos locate homes from interior photos

TechNews Editorial
TechNews EditorialOct 12, 2026 · 3 min read
Share

Why it matters

The demos show how property listing photos could help police identify a location from an interior image, though Raven's real-world performance is unclear.

The facts

  • In Graylark demos, Raven matched single interior photos to specific properties.
  • 404 Media matched demo results to Airbnb and Vrbo listings and found Zillow photos in another example.
  • Miami-Dade says its limited use of the tool has produced no significant breakthroughs or arrests.

Advertisement

Graylark Technologies has added an interior search feature to Raven, an AI tool sold to law enforcement. In company demos, Raven uses a single room photo to find a specific property. The demonstrations do not show whether it performs as quickly or reliably in routine investigations.

Graylark CEO Daniel Heinen said the feature can geolocate interior images across millions of short-term rental units worldwide. In one demo, a user uploads a social media photo of a washer and dryer, selects “Find Property: Interior and exterior property geolocation,” and chooses a search region. Within seconds, Raven shows results near Charlotte, North Carolina. The top result, “Renovated! Walk Uptown/located in 2 Hot Spots!”, matches an Airbnb listing whose photos show the same property.

A second demo starts with a bathroom photo from social media. Raven points to Montecito, California, and returns “Montecito Miramar Beach House on the Sand.” 404 Media matched that text to a Vrbo listing. The examples suggest Raven can connect a photo to other images of the same property without location metadata or an identical image online.

Other demos show homes and hotels

Raven also appears to draw on property listings beyond short-term rentals. A demo posted last year traces a room photo to a specific San Francisco address. Its returned interior photos match images on Zillow, according to 404 Media. Another video posted last week shows Raven identifying a hotel.

One video describes “City-specific property AI to determine an address,” suggesting coverage may depend on models built for particular cities. Emails obtained by 404 Media showed that the Miami-Dade Sheriff’s Office had access to a custom model trained for Miami-Dade County. The agency and the Los Angeles Police Department have bought access to Raven, previously called GeoSpy.

Graylark trained Raven on millions of images from around the world. The tool initially focused on outdoor clues such as architecture and soil. Heinen told Indicator last year that its ability to place indoor photos emerged during testing. After trying Zillow screenshots, he said, “nine out of 10 of them it would get the right city.” That account describes his test, not a measured success rate for the new property search.

Read nextSalesforce agrees to acquire Loops for Tableau analytics

Actual investigative results remain limited

Heinen said an interior match could give investigators working trafficking and exploitation cases a location lead to verify. He also said Graylark is investing heavily in 2026 in interior geolocation across U.S. hotels, short-term rentals, houses, and apartments. A LinkedIn post earlier this year showed Raven used on a video of people vandalizing a Flock camera.

Miami-Dade has reported limited results so far. Joseph R. Peguero Rivera of the sheriff’s Public Affairs Office previously told 404 Media that use of GeoSpy had been “limited and largely exploratory.” He said it had produced no significant investigative breakthroughs or arrests, and no case outcome was driven primarily by the tool.

When 404 Media first covered Raven in January 2025, it was available to essentially anyone. Some people in the company’s Discord server asked for help stalking women. A day after the outlet contacted Graylark, the company restricted access to government agencies and enterprise users. In a September X post, Heinen identified Alexandr Wang, now Meta’s head of AI, as an angel investor in Graylark. Wang did not respond to a request for comment.

Airbnb said it has no relationship with Graylark. It said its listing photos exclude location metadata, hosts can show only an approximate location, and guests receive an address after a confirmed reservation. Airbnb said its terms prohibit scraping and that it uses technical safeguards to limit automated collection. Zillow also said scraping violates its terms and that it pursues legal action when necessary. Airbnb said it is reviewing this case.

Newsletter

Get the best AI & tech news daily

A concise daily digest. Unsubscribe anytime.

We use your email only to send this newsletter.

Keep reading