Meeting Action Extractor Free

Automated Meeting Action-Item Extraction

Drowning in meeting transcripts but struggling to find the actual "to-dos"? This skill solves the problem of manual note-taking by scanning transcripts—from Otter, Fathom, Zoom, Teams, or rough notes—to produce a single, high-fidelity action items table.

What it does

The skill first cleans your transcript, stripping out timecodes and conversational filler (ums, uhs, false starts) while preserving speaker attribution. It then identifies every commitment made during the discussion and maps them into a structured format: Owner, Action, Deadline, and Outcome.

Why use this skill

• Zero Invention: Unlike general LLMs that might hallucinate owners or dates, this skill strictly flags missing information as "[needs clarification]" if it wasn't explicitly stated.
• Outcome-Oriented: It doesn't just list tasks; it captures what "done" looks like for every item.
• Developer-Friendly: Get a clean, Markdown-formatted table that you can paste directly into Jira, GitHub Issues, or Notion.
• Clean Data: It uses specific rubric-based extraction to ensure every action starts with a strong verb and maps to a specific deliverable.

Supported Inputs

Works with text exports from all major meeting platforms including Zoom, Google Meet, Microsoft Teams, Otter.ai, and Fathom, as well as unformatted manual notes.