We built StenifyAI, a lightweight tool that generates structured meeting minutes from real conversations. It works for both in-person and online meetings, and the output format changes depending on the type of meeting (product sync, client call, brainstorming, lecture, etc.). We started this during a 24-hour Cursor hackathon and spent the past month tightening the UX and accuracy.

We built it because most AI summarisation tools give generic notes that miss decisions or ownership. Our own meetings often ended with unclear action items. Instead of a one-size-fits-all summary, we built templates tuned for specific meeting formats.

Technical overview: • System audio capture for online calls • Mic capture for in-person meetings • Prompt-layer guided summaries based on meeting type • Timestamp-based parsing • Backend on Supabase, frontend in React

Current limitations: • Speaker differentiation still improving • Only basic export options • No live “assistant mode” yet — only recording/import uploads

We’d love feedback from people who run recurring product, engineering or client meetings: • Are the structures useful? • Any missing formats? • Where does it fail? • Does credit-based pricing make sense?

Link: https://stenify.ai

Still early, but happy to answer any questions about build decisions, tech stack, or process.