Summary

Add Xiaomi MiMo-V2-TTS as a first-class TTS provider alongside ElevenLabs, OpenAI, and Edge TTS.

Why

• High-quality Chinese TTS with rich style/emotion control (dialects, role-playing, singing) — fills a gap in the current provider lineup for CJK users
• Affordable pricing — currently in first-week free promo; competitive pricing going forward
• OpenAI-compatible base URL (https://api.xiaomimimo.com/v1) — familiar integration pattern
• Xiaomi already lists OpenClaw as an integration partner on their platform, so adding native TTS support strengthens the ecosystem connection

API Details

• Endpoint: POST https://api.xiaomimimo.com/v1/chat/completions
• Auth: api-key header
• Model: mimo-v2-tts
• Docs: https://platform.xiaomimimo.com/#/docs/usage-guide/speech-synthesis

Request format

{
  "model": "mimo-v2-tts",
  "messages": [
    {"role": "user", "content": "Hello, MiMo"},
    {"role": "assistant", "content": "Yes, I had a sandwich."}
  ],
  "audio": {
    "format": "wav",
    "voice": "mimo_default"
  }
}

Response format

Standard OpenAI chat completion JSON with base64-encoded audio in choices[0].message.audio.data.

Preset voices

Style control

Wrap text with style tags for expressive output:

• Emotions: happy, sad, angry (开心/悲伤/生气)
• Dialects: Northeastern, Sichuan, Cantonese, Taiwanese (东北话/四川话/粤语/台湾腔)
• Speed: faster, slower (变快/变慢)
• Roles: Sun Wukong, Lin Daiyu (孙悟空/林黛玉)
• Singing (唱歌)

Fine-grained audio tags also supported for whispers, laughter, breathing, pauses, etc.

Suggested config

{
  "messages": {
    "tts": {
      "provider": "mimo",
      "mimo": {
        "apiKey": "mimo_api_key",
        "baseUrl": "https://api.xiaomimimo.com/v1",
        "voice": "default_zh",
        "format": "wav"
      }
    }
  }
}

Implementation note

MiMo TTS uses /v1/chat/completions (not /v1/audio/speech), so it cannot simply reuse the existing OpenAI TTS provider with a custom baseUrl. A dedicated provider is needed, or it could be supported once the plugin provider system (#41496) lands.

Related

• feat(plugins): add unified provider system with capabilities #41496 — Unified plugin provider system (enables custom TTS providers via plugins)
• Feature Request: Add xAI (Grok) TTS as a native TTS provider #48454 — xAI TTS provider
• feat(tts): add Deepgram Aura as first-class TTS provider #43579 — Deepgram Aura TTS provider
• feat(tts): add MiniMax TTS provider support #49894 / feat(tts): add MiniMax as a TTS provider #44716 — MiniMax TTS provider
• Add Fish Audio as TTS provider #52019 — Fish Audio TTS provider