Welcome back!

Last month we published the 100th Bellingcat challenge. To celebrate, we are launching a brand-new website! This is an exciting milestone for the newsletter and we thank you for your support. Just under 10,000 people now receive the challenges every month. 

Here’s to the next 100 challenges! 

But let’s not get ahead of ourselves just yet. First, we have five fresh mysteries for you to decode, as well as sharing the answers to last month’s challenges.

Bellingcat is the only investigative organization where the collective builds society's capacity to find the truth. We turn obsessive curiosity into a teachable research skill, build the public community space to practice it, and continually innovate solutions to drive discovery forward. You can support our work here: https://www.bellingcat.com/donate/

When geolocating an image, the key is to look for features that are semi-permanent, things that are unlikely to change from day to day; like buildings, street signs or roads. But what if you’re trying to find the location of a photo taken in areas with little to no infrastructure? What natural features can you use to geolocate images?

There are many examples of this to be found in Bellingcat investigations. A 2025 story about wildfires in a national park in Namibia includes a geolocation that uses a termite mound to pinpoint the location. Termite mounds can take years to build and can last for hundreds of years. Therefore, they are semi-permanent features! 

Another example can be found in this investigation which includes a very impressive geolocation in the middle of the Libyan desert that utilised patterns in the sand visible on satellite images and a mountaineering tool called Peakvisor. You can read more about Peakvisor in the Bellingcat Toolkit, it might come in handy…

So, think you’re ready? Find the latest challenges here. 

If you get stuck, make sure to join our Discord server to discuss the challenges with others!

SPOILER ALERT: From this point on, we’ll be discussing last month’s challenges. If you still plan to complete them, you may want to skip this section.

Here are the answers to the “Editing Errors” challenges:

For the ‘Corn Conundrum’ challenge, Faaiz K blogged about how they quickly spotted the image had been altered. They used Forensically, a collection of web-based image and forensic tools that can identify altered images, to check if details have been tampered with. The name of the company on the van stood out, which suggests that it might not be the original name. 

But there was also an address on the car which hadn’t been altered. 

Checking Google Maps and other map providers, did not reveal any details about the business at this location. But just before Faaiz started to consider other approaches, they spotted the same van on Google Streetview. It was parked just metres away from the address and included the original text written on the side of the door. 

Nakanishi tried several approaches to solve the ‘Losing Focus’ challenge. After reverse image searching and upscaling the image without success, they took a closer look at the filename. The 30x40 image size stood out, suggesting it was a resized version of the original image. By manually modifying the URL and removing 30x40, they were able to access a higher-resolution version. In this version, the name of the coffee shop was clearly readable.

‘Lost in Translation’ proved easy enough to solve using reverse image search or, like Nakanishi did, using ChatGPT. The text had initially been altered using ChatGPT, but that same program recognised that what remained was from a Code of Procedure in the Portuguese Administrative Courts. Matching the dates in the text to a section of this document allowed Nakanishi to find that the last word listed was “verdade” and thus, the right answer.

‘Pixel Perfect’ was solved by Oliver Parker who posted about his method on LinkedIn. The advertisement banner in the photo proved to be a dead end, so they opted to perform a reverse image search to see what they could find. TinEye and Yandex reverse image search came back with nothing, and Google initially also struggled to recognise features in the picture. But after some tinkering, it eventually suggested a cathedral in Albania. Using Google Maps Oliver could verify the location near the church and find the exact spot that the image was taken at.

Despite the photo in ‘Planted Evidence’ being modified by AI, reverse image search tools still recognised that it was likely taken in Venice. Yet further details were more challenging to find. Many participants shared that they spent hours looking at reverse image search results or scrolling on Google Maps to no avail. But Nick Xifaras found a solution by looking at the carving stone roundel set into the wall above the door. In Venice these are called patere. Nick found a publicly available spreadsheet full of documented pantere around Venice on the veniceprojectcenter.org website. By applying filters to narrow the number of pantere to a manageable set, he identified one that appeared to match the image, listed as "gryphon above hare/rabbit." It also included an address. A quick check on Google Maps confirmed that this was indeed the correct location.

This month, you can also find video walkthroughs of some of the challenges. Kbaa published a few on his YouTube channel, go check it out!

That’s it for this month’s Bellingcat Challenge Newsletter. We’d love to hear your feedback on the challenges and the new website. Join us on Discord and let us know if you have ideas for future challenges.

Before we go, here are some links to other current projects from Bellingcat: