groq-transcribe

使用Groq的Whisper API将音频文件转录为文本。当您收到音频/语音消息、需要将语音转换为文字，或处理各类音频文件（.ogg、.mp3、.wav、.m4a、.flac、.webm）时，此方法可助您事半功倍。

SKILL.md

--- frontmatter

name: groq-transcribe
description: Transcribe audio files to text using Groq's Whisper API. Use when you receive audio/voice messages, need to convert speech to text, or process audio files (.ogg, .mp3, .wav, .m4a, .flac, .webm).

Groq Transcription

Fast speech-to-text via Groq API (~2 seconds for voice messages).

Setup

Get a free API key at console.groq.com.

Add to OpenClaw config:

json

{
  "skills": {
    "entries": {
      "groq-transcribe": {
        "enabled": true,
        "env": {
          "GROQ_API_KEY": "gsk_your_key_here"
        }
      }
    }
  }
}

Or set GROQ_API_KEY environment variable.

Usage

bash

{baseDir}/scripts/transcribe.py <audio_file>

Output: plain text transcription to stdout.

Limits

Tier	Requests	File size	Cost
Free	20/min, 2000/day	25 MB	$0
Dev	100/min, 50k/day	100 MB	$0.04/hr audio