Transcribe audio via Groq API (~10x cheaper than OpenAI API)

Transcribe audio via Groq Automatic Speech Recognition (ASR) Models (Whisper).

Security Scan

The skill's requests and instructions match its stated purpose: it needs curl and a GROQ_API_KEY and simply uploads an audio file to Groq's transcription endpoint and writes the transcript locally.

Purpose & Capability

Name/description, required binary (curl), and required env var (GROQ_API_KEY) align with a simple ASR transcription helper. There are no unrelated credentials, binaries, or config paths requested.

Instruction Scope

SKILL.md and the included script only read the provided audio file, the GROQ_API_KEY environment variable, and write the transcript output. The script POSTs the file to https://api.groq.com/openai/v1/audio/transcriptions (the documented Groq endpoint). It does not read other system files or exfiltrate data to unexpected endpoints.

Install Mechanism

No install spec; this is instruction-only with an included shell script. No downloads, package installs, or archive extraction are performed by the skill itself.

Credentials

Only GROQ_API_KEY (primary credential) is required, which is appropriate for calling Groq's API. The README shows an optional place to store the key (~/.openclaw/openclaw.json) — this is a convenience tip, not a hidden requirement; users should still secure that file if used.

Persistence & Privilege

always is false and the skill does not request or modify other skills or system-wide config. It can be invoked autonomously (platform default), which is expected for a callable skill but not otherwise problematic here.

Assessment

This skill appears to do exactly what it says: it uploads audio to Groq's transcription endpoint and saves the transcript locally. Before installing, confirm you trust Groq to process your audio (sensitive audio will be sent off-host). Protect your GROQ_API_KEY (store it in a secure place, use least-privilege/rotating keys if possible) and be cautious about putting the key into files like ~/.openclaw/openclaw.json unless that file is secured. If you do not want the agent to call this skill autonomously, restrict autonomous invocation in your agent settings. Otherwise this skill is coherent and proportional for the stated purpose.

Groq Whisper API (curl)

Transcribe an audio file via Groq’s OpenAI-compatible /openai/v1/audio/transcriptions endpoint.

Quick start

{baseDir}/scripts/transcribe.sh /path/to/audio.m4a

Defaults:

• Model: whisper-large-v3-turbo
• Output: <input>.txt

Useful flags

{baseDir}/scripts/transcribe.sh /path/to/audio.ogg --model whisper-large-v3 --out /tmp/transcript.txt
{baseDir}/scripts/transcribe.sh /path/to/audio.m4a --language en
{baseDir}/scripts/transcribe.sh /path/to/audio.m4a --prompt "Speaker names: Peter, Daniel"
{baseDir}/scripts/transcribe.sh /path/to/audio.m4a --json --out /tmp/transcript.json

API key

Set GROQ_API_KEY, or configure it in ~/.openclaw/openclaw.json:

{
  skills: {
    "groq-whisper-api": {
      apiKey: "GROQ_KEY_HERE",
    },
  },
}