Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
macOS CLI tool for transcribing audio and video files using local Whisper models or Whisnap Cloud. Requires the Whisnap desktop app with at least one model downloaded. Outputs plain text or structured JSON with timestamps.
Local model transcription keeps audio private on-device, while the cloud option is available when speed matters, all from a single CLI command.
When to use it
- Transcribe recorded meetings to text
- Extract captions from lecture recordings
- Script batch transcription of interview files
- Get timestamped segments from podcast audio
- Transcribe screen recordings locally without uploading
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: transcribe a meeting recording and get timestamped JSON output
- 1Run `whisnap meeting.mp4 --json` to start transcription using the default local model
- 2Whisnap reads the file and processes it through the downloaded Whisper model
- 3Structured JSON is written to stdout with full text, per-segment timestamps, model name, and processing time
- 4Pipe stdout to a file or downstream tool for further processing
JSON object with `text`, `segments` array (each with `start_ms`, `end_ms`, `text`), `model`, `backend`, and `processing_time_ms`
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.