transcribe

name	transcribe
description	Convert audio or video speech into structured text with timestamps, speaker hints, and downstream-ready summaries.

Transcribe

Use this skill when Thomas needs reliable text from audio or video before analysis or editing work continues.

Workflow

1. Confirm the source file, language, and expected output shape.
2. Transcribe with timestamps and speaker labeling when available.
3. Normalize names, obvious transcription errors, and action items carefully.
4. Preserve uncertainty markers when speech is ambiguous.
5. Hand off both the transcript and any concise summary the task needs.

Rules

• Do not invent speaker certainty that the audio does not support.
• Keep timestamps when later editing or clipping depends on them.