A transcript is a record. Minutes are a promise

Description and use

A local, automated meeting recording system using Whisper and LM Studio generates structured minutes with action items and owners, ensuring accountability and easy retrieval.

Problem

Meeting participants often forget action items and decisions, leading to missed deadlines and accountability gaps.

Outcome

The system enabled the author to reliably capture commitments, retrieve them quickly, and hold participants accountable without relying on external services or subscriptions.

Workflow

• 1. Start the local recording app before the meeting.
• 2. Record mic and system audio from the chosen platform.
• 3. Whisper transcribes the audio locally.
• 4. LM Studio processes the transcript to produce Summary, Key Decisions, Action Items, Open Questions.
• 5. Action items include owner and due date when specified.
• 6. Save the output as Markdown/PDF in a dedicated folder.
• 7. Refer to the file for follow‑ups and accountability checks.

Evidence

I run the thing I built for every meeting I'm in. It records mic and system audio (Teams, Zoom, Meet, or a room full of people), transcribes it locally with Whisper, then a small language model I run in LM Studio writes it up as Summary, Key Decisions, Action Items, Open Questions.

Source

Everything runs on my own hardware. Windows 10/11, Python 3.10+, and LM Studio with any small model loaded. It works on a plain CPU — a GPU just makes transcription faster.

Source