Teams sit through hour-long meetings then forget action items and decisions.
A Python service that takes a meeting transcript (or audio), splits by speaker, summarizes decisions, extracts action items with owners and deadlines, and emails a digest.
Speaker diarization accuracy and detecting action items buried in casual conversation.
Whisper transcription, prompt chaining for action item extraction, and async task queues.
Slack/Teams integration and a dashboard of recurring action items.
Log in to rate and review this project.
Extracting action items and owners from a raw transcript is genuinely useful. The summary is concise and the action items are accurate. Speaker attribution works well even with overlapping dialogue.