Rosetta Map lets you read one city through another. Pick a base city and a labels city, and every neighborhood is relabeled with its closest analog.
GPT-5.4 reviews neighborhoods for recognizability, then scores each one across dimensions like density, cost, nightlife, transit, and cultural significance.
For each pair of cities, GPT-5.4 runs bidirectional matching: for every neighborhood in City A, it finds the closest analogs in City B, and vice versa. Each match gets a similarity score and a rationale.
Final pairings are computed using the Hungarian algorithm (maximum weight bipartite matching), not a greedy approach, so the global assignment is optimal. Street name analogs are matched separately using the same process.
Made by Riley Walz