Short version: the photos are real, the wings are real, and an AI language model helped write and organize the posts. Here’s the actual process, no mystery about it.
How a post gets made
Every post starts from a real photo, pulled from Google Photos (the only photo source this site uses — read-only access, nothing automated or continuous yet). GPS coordinates embedded in each photo’s EXIF data get extracted and run through offline reverse geocoding to figure out roughly where it was taken. Photos taken close together in time and location get grouped into a single post.
From there, an AI assistant does the writing: researching a restaurant when there’s enough to go on, drafting the post in the site’s usual tone, and tagging it with location and category data. A human reviews and directs the whole thing — nothing gets published unread.
The honesty rules
- Restaurant posts never fabricate a name or a link. If a venue can’t be confidently identified from the photo and location data, the post says so honestly instead of guessing.
- Home-cooked posts are allowed one small invented detail — something plausible about that particular cooking session — but it’s always written to read as an obvious guess, never presented as fact.
- Home locations stay at the town level, never an address. Precise GPS coordinates exist in this site’s data for internal use (nearby-post features, maps for actual restaurants) but are never displayed for home-cooked posts.
If something on this site looks wrong — a misidentified restaurant, a bad guess, anything — that’s a mistake worth fixing, not a feature. Feel free to assume the wings themselves are the only part that’s never exaggerated.