Echos — AI Meeting & Voice-Memo Assistant

Meetings and voice memos are easy to record but hard to act on. Important context often gets buried in raw audio that nobody re-listens to, and decisions or action items get lost in the process. Echos solves this by turning any spoken memo into a clear, structured written record. Users can record audio directly in the browser or upload an existing file. The audio is sent to AssemblyAI, which transcribes it, generates an AI summary, extracts key highlighted phrases, and runs sentiment analysis across the conversation. The result is displayed in a clean, visual interface: a full transcript with a confidence score, a concise summary, a set of key phrase chips, and a color-coded sentiment breakdown showing the balance of positive, neutral, and negative moments in the recording. Built as a single self-contained web app with no backend required, Echos uses the browser's native MediaRecorder API for cross-browser audio capture, ensuring it works reliably regardless of platform. This was built solo, end-to-end, from problem framing to a fully working, API-integrated product — including the full AssemblyAI transcription, summarization, highlights, and sentiment pipeline.