By 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
- Confirm the source file, language, and expected output shape.
- Transcribe with timestamps and speaker labeling when available.
- Normalize names, obvious transcription errors, and action items carefully.
- Preserve uncertainty markers when speech is ambiguous.
- 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.