Skip to content
TakoAPI
Back to skills

audio-transcribe

by aktheknight

Media & Streaming
817Downloads reported by clawskills.sh
0Views
0 likesSign in to like this skill
—Not yet rated
Your rating

Brief

Auto-transcribe voice messages using faster-whisper (local, no API key needed).

Description

What This Skill Does

Transcribes audio files locally using faster-whisper, a CPU-compatible Whisper implementation. No API key or internet connection is required for transcription. Clawdbot can auto-transcribe incoming voice messages when the skill is enabled.

Runs entirely on-device, so audio never leaves the machine and there are no per-request costs or usage limits.

When to Use It

  • Transcribing voice notes sent over messaging apps
  • Converting recorded calls or meetings to text
  • Archiving voice memos as searchable transcripts
  • Captioning audio files for accessibility reviews
  • Processing voice messages without uploading to external services

Example Workflow

INPUT

User asks: transcribe this voice message at /tmp/message.ogg

AGENT
  1. 1

    Locate the audio file at the specified path

  2. 2

    Run transcribe.py with the file path as argument

  3. 3

    faster-whisper loads the configured model (small by default)

  4. 4

    Model processes the audio locally and outputs a text transcript

OUTPUT

Plain text transcript of the audio content

Installation

clawhub install aktheknight/audio-transcribe

Run the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.

Links