Drop your bank & utility CSVs, go to bed 🦥 By morning, Napper's local TabPFN-3.5 foundation model delivers an audio brief. It caught a +706% water burst on Day 1 and a gym hike.

Developer Ansuj Kumar Meher built Napper, a fully offline agent that ingests bank and utility CSVs and uses a local TabPFN-3.5 foundation model to generate an audio brief overnight. In its first day of operation the agent flagged a +706% water-usage spike and a gym membership price hike, running at zero API cost.

I built an agent that works while I sleep. It found a pipe leak, a price hike, and a subscription I forgot I had. Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady Submission 🦥 Ansuj Kumar Meher Ansuj Kumar Meher Ansuj Kumar Meher Follow Oct 25 I built an agent that works while I sleep. It found a pipe leak, a price hike, and a subscription I forgot I had. devchallenge hf26challenge opensource ai 1 reaction Add Comment 7 min read