Kai Minimax Tts

Generate voice audio and transcribe speech using MiniMax TTS API. Use when responding with voice or transcribing audio files.

Security Scan

The skill's code, declared requirements, and runtime instructions match its stated purpose (calling MiniMax TTS and running a local transcription), and there are no disproportionate permissions or surprising behaviors.

Purpose & Capability

Name/description, required env var (MINIMAX_API_KEY), and required binaries (curl, xxd, whisper) align with a TTS + transcription skill. The script calls a MiniMax API endpoint and uses whisper locally for transcription—these requirements are expected for this functionality.

Instruction Scope

SKILL.md and the included script stick to the stated tasks: generating audio via the MiniMax API and transcribing via a local whisper binary. The script only reads the declared MINIMAX_API_KEY and optional KAI_* env vars, writes outputs into a per-skill workspace, and does not attempt to read unrelated files or credentials.

Install Mechanism

This is an instruction-only skill with an included shell script; there is no install spec that downloads or executes remote code. No high-risk install behavior detected.

Credentials

Only MINIMAX_API_KEY is required (plus optional workspace and voice ID env vars). That single credential is appropriate for calling the external MiniMax API; no unrelated secrets or multiple service credentials are requested.

Persistence & Privilege

The skill does not request always:true and does not change other skills or system-wide settings. It creates and writes files within its workspace only, which is normal for this type of skill.

Assessment

This skill appears to do exactly what it says: it sends the text you provide to MiniMax's API (https://api-uw.minimax.io) using the MINIMAX_API_KEY and writes an MP3 and a transcript into a workspace directory. Before installing: 1) Ensure the MINIMAX_API_KEY you provide is only for the intended MiniMax service and that you trust that endpoint. 2) Confirm the whisper binary on your system is the implementation you expect (local vs. a wrapper that might upload audio). 3) Be aware output files overwrite fixed names (Kai.mp3 and latest_from_blaze.txt) in the skill workspace; customize KAI_MINIMAX_WORKSPACE if you want a different path. 4) If you need stricter isolation, run the script in a controlled environment since it performs network I/O (to the TTS API) and writes files locally.

Generate voice audio using MiniMax TTS.

Setup

Add to ~/.openclaw/openclaw.json:

{
  "skills": {
    "entries": {
      "kai-minimax-tts": {
        "enabled": true,
        "env": {
          "MINIMAX_API_KEY": "your_key_here"
        }
      }
    }
  }
}

Usage

Generate voice (English):

bash {baseDir}/scripts/kai_tts.sh --speak "Hello" en

Generate voice (Spanish):

bash {baseDir}/scripts/kai_tts.sh --speak "Hola" es

Transcribe audio:

bash {baseDir}/scripts/kai_tts.sh --transcribe /path/to/audio.ogg

Output Files

• Voice: $WORKSPACE/Kai.mp3
• Transcript: $WORKSPACE/latest_from_blaze.txt

Customization

Set custom workspace:

"env": { "KAI_MINIMAX_WORKSPACE": "/custom/path" }

Set custom voice IDs:

"env": { 
  "KAI_ENGLISH_VOICE_ID": "voice_id",
  "KAI_SPANISH_VOICE_ID": "voice_id"
}