MiniMax TTS Generator

Text-to-speech (TTS) generation using MiniMax API. Converts text into natural-sounding speech with support for multiple voices, adjustable speed and pitch, a...

Security Scan

The skill mostly does what it says (TTS via MiniMax) and only asks for the expected API key, but there are a few mismatches and small risks (missing declared runtime dependency, an undocumented overrideable API host, and use of subprocess/ffmpeg) that users should be aware of before installing.

Purpose & Capability

Name/description match the code: the script sends text to MiniMax endpoints and returns/writes audio. Required binary python3 and MINIMAX_API_KEY are appropriate for a TTS client. The network endpoints in code point to api.minimaxi.com which aligns with the stated MiniMax purpose.

Instruction Scope

SKILL.md instructs running the included Python script and storing the API key in env or openclaw config — this matches the code. The runtime does read the user's openclaw config (~/.openclaw/openclaw.json) and writes audio to ~/.openclaw/workspace/tmp; both are documented in SKILL.md. The script uses subprocess.run to call ffmpeg when concatenating segments; the SKILL.md examples show multi-segment usage but do not declare ffmpeg as a required binary. Also the code allows overriding the API host via MINIMAX_API_HOST (which can redirect network calls) — this override is not declared as a required env var in the registry metadata.

Install Mechanism

This is instruction-only / contained code included in the skill bundle with no network-download install step. No external archives or installers are fetched during install, which lowers supply-chain risk.

Credentials

The only declared required credential is MINIMAX_API_KEY (primaryEnv) which is proportionate for a TTS API client. The code also reads MINIMAX_API_HOST if set (not declared in requires.env) and the openclaw config file; both are reasonable but the undocumented MINIMAX_API_HOST allows redirecting requests to an arbitrary host and should be noted.

Persistence & Privilege

The skill does not request always:true, does not demand extra system privileges, and confines storage to ~/.openclaw/workspace/tmp and its own config entry in ~/.openclaw/openclaw.json. It does not modify other skills' settings or system-wide configs beyond its own entry.

What to consider before installing

This skill appears to implement a legitimate MiniMax TTS client and only requires your MiniMax API key. Before installing: (1) Review and confirm you are comfortable providing MINIMAX_API_KEY (the script sends it as a Bearer token to the API). (2) Install ffmpeg if you plan to use multi-segment output (the script calls ffmpeg but ffmpeg is not listed as a required binary). (3) Be aware the code will read ~/.openclaw/openclaw.json and write audio files to ~/.openclaw/workspace/tmp (check those paths and permissions). (4) The script supports an environment override MINIMAX_API_HOST — don't set this to an untrusted host (an attacker-controlled host could capture your API key). (5) If you do not fully trust the skill source, review the included minimax_tts.py before running it (it performs network requests and executes ffmpeg via subprocess). If these points are acceptable or fixed (declare ffmpeg as required, document MINIMAX_API_HOST), the skill is coherent with its stated purpose.

MiniMax TTS (Text-to-Speech)

⚠️ Requires a MiniMax API Key with TTS/Speech access

• Coding Plan key (sk-cp-xxx) → 不支持 TTS，仅限 M2.7 聊天模型
• 通用 API Key → 请在 MiniMax 开放平台 确认您的订阅是否包含语音合成
• TTS 独立套餐: 访问 MiniMax 语音合成产品

Text-to-speech synthesis using MiniMax's speech synthesis API. Supports 40+ languages, multiple voices, adjustable speed/pitch, and various audio formats.

---

Setup

1. Configure API Key

openclaw config set skills.entries.minimax-tts.apiKey "sk-your-key"

Or add to openclaw.json skills entries:

{
  "skills": {
    "entries": {
      "minimax-tts": {
        "apiKey": "sk-your-key"
      }
    }
  }
}

2. Dependencies

pip install requests

---

Architecture

~/.openclaw/workspace/skills/minimax-tts/
├── SKILL.md
├── _meta.json
└── scripts/
    └── minimax_tts.py

---

Usage

From terminal

# Basic TTS - text to speech
MINIMAX_API_KEY="sk-your-key" python3 scripts/minimax_tts.py "你好，欢迎使用 MiniMax 语音合成"

# With voice selection
MINIMAX_API_KEY="sk-your-key" python3 scripts/minimax_tts.py "Hello world" --voice female-shaonv

# Adjust speed and pitch
MINIMAX_API_KEY="sk-your-key" python3 scripts/minimax_tts.py "快速播报新闻" --speed 1.5 --pitch 2

# Save to file
MINIMAX_API_KEY="sk-your-key" python3 scripts/minimax_tts.py "输出到文件" --output speech.mp3

# List available voices
MINIMAX_API_KEY="sk-your-key" python3 scripts/minimax_tts.py --list-voices

# Multi-segment (audiobook/podcast with multiple voices)
MINIMAX_API_KEY="sk-your-key" python3 scripts/minimax_tts.py --segments segments.json --output combined.mp3

From code

from minimax_tts import generate_speech

# Basic TTS
result = generate_speech(text="你好世界", voice="female-shaonv", output_path="hello.mp3")

# With options
result = generate_speech(
    text="这是一段测试语音",
    model="speech-2.8-hd",
    voice="female-tianmei",
    speed=1.0,
    pitch=0,
    audio_format="mp3",
    output_path="test.mp3"
)

---

Tool Definition

Name: minimax_tts

Input Schema:

{
  "type": "object",
  "properties": {
    "text": {
      "type": "string",
      "description": "Text to convert to speech (max 10000 chars per request)"
    },
    "model": {
      "type": "string",
      "enum": ["speech-2.8-hd", "speech-2.6-hd", "speech-02-hd"],
      "default": "speech-2.8-hd",
      "description": "TTS model: speech-2.8-hd (recommended, high quality with emotion control), speech-2.6-hd (Coding Plan NOT supported), speech-02-hd (compact HD voice)"
    },
    "voice": {
      "type": "string",
      "default": "female-tianmei",
      "description": "Voice ID: female-tianmei, female-shaonv, female-yujie, female-chengshu, male-qn-qingse, male-qn-jingying, male-qn-badao, male-qn-daxuesheng"
    },
    "speed": {
      "type": "number",
      "default": 1.0,
      "minimum": 0.5,
      "maximum": 2.0,
      "description": "Speech speed, range 0.5-2.0, default 1.0"
    },
    "pitch": {
      "type": "number",
      "default": 0,
      "minimum": -12,
      "maximum": 12,
      "description": "Voice pitch adjustment, range -12 to 12, default 0"
    },
    "audio_format": {
      "type": "string",
      "enum": ["mp3", "wav", "pcm"],
      "default": "mp3",
      "description": "Output audio format"
    },
    "output_path": {
      "type": "string",
      "description": "Save audio to this file path. Default: saved to ~/.openclaw/workspace/tmp/tts_<timestamp>.mp3"
    }
  },
  "required": ["text"]
}

Output: JSON with success status and audio file path

---

Common Voice IDs

voice_id	Description	Gender
female-tianmei	甜美女声	Female
female-shaonv	少女音	Female
female-yujie	御姐音色	Female
female-chengshu	成熟女性	Female
male-qn-qingse	青年男声（青涩）	Male
male-qn-jingying	青年男声（精英）	Male
male-qn-badao	青年男声（霸道）	Male
male-qn-daxuesheng	青年大学生	Male

---

Multi-Segment Generation (Audiobooks / Podcasts)

For multi-character content, create a segments.json file and use --segments:

python3 scripts/minimax_tts.py --segments segments.json --output combined.mp3

segments.json Format

[
  {
    "text": "Morning sunlight streamed into the room.",
    "voice": "female-tianmei",
    "speed": 1.0
  },
  {
    "text": "Welcome to today's episode.",
    "voice": "male-qn-qingse",
    "speed": 1.0
  }
]

---

Error Codes

Code	Meaning
0	Success
1002	Rate limit - try again later
1004	Auth failed - check API Key
1008	Insufficient balance
1026	Content violation
2013	Parameter error
2049	Invalid API Key