Summary

Feature Description

MiniMax Token Plan includes TTS HD capability (speech-2.8-hd / speech-2.6-hd / speech-02-hd),
but the OpenClaw MiniMax plugin only exposes text (M2.7) and image generation (image-01).
The built-in TTS module falls back to Microsoft Edge TTS instead of MiniMax TTS HD.

Expected Behavior

When using the tts tool with MiniMax OAuth authentication, it should:

1. Use the MiniMax TTS HD models available in the user's Token Plan
2. Consume the TTS HD quota from the Token Plan (not external paid TTS)
3. Support voice options like emotion control and voice selection

Current Workaround

The vm0-ai/vm0-skills@minimax skill can call MiniMax TTS API directly, but it requires
a separate API key and doesn't integrate with Token Plan quotas.

Additional Context

• MiniMax documentation: https://platform.minimaxi.com/docs/token-plan/faq
• TTS HD is listed as one of the core non-text models in Token Plan
• This affects all users who want to use voice synthesis with their MiniMax subscription

Problem to solve

Problem to Solve

1. Wasted subscription benefit: Users with MiniMax Token Plan have TTS HD quota, but cannot use it through OpenClaw — the built-in TTS completely ignores MiniMax and falls back to Edge TTS, which produces empty audio.
2. Forced to pay twice: To use MiniMax TTS, users must pay for a separate API key outside their Token Plan, even though TTS HD is already included in their subscription.
3. Inconsistent experience: Text and image generation correctly use MiniMax Token Plan, but TTS silently fails and uses an unrelated Microsoft TTS service.

Proposed solution

Proposed Solution

Extend the MiniMax plugin to include TTS HD support, similar to how text and image
generation are already integrated. The TTS should use the user's Token Plan

Alternatives considered

No response

Impact

Impact

• Affected users: All OpenClaw users with MiniMax Token Plan subscription who want to use voice synthesis (primarily Feishu, Discord, and other chat channel users)
• Severity: Annoying / Blocks workflow — TTS is completely non-functional for MiniMax users; the tool returns an error instead of providing the expected voice output
• Frequency: Always — every time a user tries to use the tts tool with MiniMax OAuth, it fails and falls back to Edge TTS (which also fails)
• Consequence: Users must manually work around by using third-party TTS services or giving up on voice synthesis entirely. This breaks the expected "all-in-one" experience where the Token Plan should cover all modalities.

Evidence/examples

No response

Additional information

No response