openclaw - 💡(How to fix) Fix [Feature]: Add MiniMax Speech as TTS Provider [1 comments, 2 participants]

Summary

Add MiniMax Speech-02 as a fourth TTS provider option alongside ElevenLabs, OpenAI, and Edge.

Problem to solve

OpenClaw currently only supports three TTS providers: ElevenLabs, OpenAI, and Edge. Many users who already use MiniMax for LLM capabilities (via OAuth or API) want to use MiniMax Speech-02 for text-to-speech, but it's not available. The existing TTS options either require paid accounts (ElevenLabs, OpenAI) or have limited voice quality (Edge). MiniMax Speech-02 offers significantly better Chinese TTS quality (WER 2.25% vs ElevenLabs 16%), lower cost (1/4 of ElevenLabs), and already integrates with many platforms like LiveKit, Vapi, and Pipecat.

Proposed solution

Add "minimax" as a TtsProvider option in the OpenClaw config schema:

1. Add to TtsProvider type: "minimax"
2. Implement MiniMax TTS adapter using MiniMax Text-to-Speech API (speech-02, speech-02-hd models)
3. Support configuration via messages.tts.provider: "minimax" with model options (speech-02, speech-02-hd, speech-02.6)
4. Support voice selection, speed control, and emotion control parameters
5. Leverage existing MiniMax OAuth integration for authentication

Configuration example: { "messages": { "tts": { "provider": "minimax", "minimax": { "model": "speech-02-hd", "voiceId": "male-qn-qingse", "speed": 1.0, "emotion": "chat" } } } }

Alternatives considered

1. Using ElevenLabs or OpenAI TTS - requires separate paid accounts, higher cost, worse Chinese voice quality
2. Using Edge TTS - free but limited voice options and lower quality
3. Community-developed skills (minimax-tts) - not officially supported, may have stability issues
4. Using MiniMax API directly via custom scripts - requires manual setup, not integrated into OpenClaw's TTS pipeline

Impact

• Affected users: OpenClaw users who use MiniMax for LLM and want unified TTS; Chinese-speaking users who need high-quality Chinese TTS
• Severity: Medium-High - blocks adoption of MiniMax as complete AI platform solution
• Frequency: Constant - every user who wants MiniMax TTS
• Consequences: Users must maintain separate TTS subscriptions (ElevenLabs/OpenAI) alongside their MiniMax subscription, increasing costs and complexity

Evidence/examples

• MiniMax Speech-02 benchmarks: https://36kr.com/p/3293468160493829
• MiniMax TTS API docs: https://platform.minimax.io/docs/api-reference/speech-t2a-intro
• MiniMax Speech 2.6 announcement: https://www.minimax.io/news/minimax-speech-26
• Previous feature requests: #20264, #20021, #27166 (all closed without resolution)
• OpenClaw already supports MiniMax as LLM provider - adding TTS completes the integration

Additional information

OpenClaw already has MiniMax OAuth integration, which means users are already authenticated with MiniMax. This makes adding MiniMax TTS a natural extension that would require minimal additional authentication setup. Many community members have requested this feature (evidenced by multiple closed issues/PRs), showing clear user demand.