Devlog #15 ->

Part 1 (still working on this one!)

Finally started actually tackling the AI usage limit problem from devlog #2. Since everyone using the app shares the same free-tier Gemini key, scans would just randomly fail with no real pattern whenever enough people were hitting it at once. Started on the backend side first: instead of letting every incoming scan hit Gemini immediately, I added a little queue that only lets 2 requests be in-flight to Gemini at a time, and everything else waits its turn. Next step is getting it to actually retry instead of just failing after the wait, but that’s for the next devlog.