Skip to content
TakoAPI
Back to skills

voice-to-text

by vae999

Media & Streaming
—Not yet rated
Your rating

Brief

Convert voice messages and audio files to text using Vosk offline speech recognition.

Description

What This Skill Does

Converts voice messages and audio files to text using Vosk, an offline speech recognition toolkit. Supports MP3, WAV, M4A, OGG, FLAC, AAC, and WEBM formats across multiple languages. No internet connection is required once a language model is downloaded.

Runs entirely offline after model download, so audio files never leave the local machine.

When to Use It

  • Transcribing WeChat or Telegram voice messages to readable text
  • Converting recorded voice memos into written notes
  • Processing OGG or M4A audio clips without uploading to a cloud service
  • Transcribing multi-language audio by switching Vosk models
  • Archiving meeting or interview recordings as searchable text

Example Workflow

INPUT

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

AGENT
  1. 1

    Accept the audio file path from the user

  2. 2

    Run transcribe.py with the provided file path

  3. 3

    ffmpeg converts the audio to 16kHz mono WAV internally

  4. 4

    Vosk model processes the WAV file and generates a transcript

  5. 5

    Return the full transcription to the user

OUTPUT

Plain text transcription of the voice message

Requirements

  • Vosk language model file downloaded to ~/.vosk/models/ from https://alphacephei.com/vosk/models

Installation

clawhub install vae999/voice-to-text

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

Links