Tried it out just now — really smooth experience. Feels way simpler than messing with ChatGPT image plugins or figuring out prompts. I am tech guy so it is easy to create a prompt, but others might not. The styles are actually pretty polished too. This seems like a good fit for people who just want a nice-looking, personalized image without jumping through hoops.
Nice. Clean and lightweight compared to full routing stacks like OSRM or Graphhopper, which can be a bit much for smaller/local tasks. Curious why the README calls it an "approximation" though — Dijkstra gives exact shortest paths unless you're simplifying inputs. Maybe that's referring to free-flow assumptions or limitations in the underlying network data? Still, cool to see a focused US-only tool.
