← 返回 Skills 市场
128
总下载
0
收藏
0
当前安装
8
版本数
在 OpenClaw 中安装
/install mimotts25-plus
功能描述
🚀 小米 MiMo TTS 2.5 Plus — 基于 mimotts25 深度优化增强。升级到 v2.5-tts 模型，新增声音设计（文字定制音色）、声音克隆（音频复刻）、导演模式等高级功能。支持 v2.5 新音色：冰糖、茉莉、苏打、白桦等。
安全使用建议
This skill's code and instructions implement a MiMo TTS client (preset voices, voice design, voice cloning) and do require an API key (MIMO_API_KEY) — despite the registry header claiming none. Before installing, verify the following: (1) the owner/source and homepage are missing — only install if you trust the publisher or have inspected the code; (2) provide a MiMo-specific API key, not a high-privilege or unrelated key (the scripts will accept OPENAI_API_KEY as a fallback, which could accidentally expose an unrelated key to the MiMo endpoint); (3) voice cloning uploads local audio (base64) to the configured API endpoint — only use samples you’re allowed to upload; (4) confirm API_BASE (default token-plan-cn.xiaomimimo.com) is the expected endpoint for your account/region; and (5) ask the publisher to fix the manifest inconsistency that omits MIMO_API_KEY from registry metadata. If you cannot verify the source, run the scripts in an isolated environment and inspect/replace API endpoint and keys before use.
能力标签
能力评估
Purpose & Capability
The name/description (MiMo TTS 2.5 Plus) align with the provided scripts: they call a MiMo TTS API, support preset voices, voice design, and voice cloning. Requiring an API key is appropriate. However, the registry metadata at the top declares no required environment variables / primary credential, while the SKILL.md and all scripts clearly require MIMO_API_KEY (with OPENAI_API_KEY as a fallback). This metadata mismatch is an incoherence to be resolved.
Instruction Scope
Runtime instructions and scripts stay within the TTS domain: they build messages, base64-encode audio samples for voice cloning, and POST to the configured MiMo API endpoint. They do not attempt to read unrelated system config or secrets beyond environment variables and user-supplied audio files. The documented user/message role conventions are internal to the TTS API and consistent across files.
Install Mechanism
No install spec or remote downloads are included — this is instruction+script-only. All included code is local and will not fetch arbitrary archives during install, which lowers supply-chain risk. The scripts do use standard Python stdlib and optionally the openai package if present (fallback to urllib).
Credentials
The scripts legitimately require an API key (MIMO_API_KEY) and accept OPENAI_API_KEY as a fallback for convenience; this is proportionate to contacting the MiMo service. The concern is the manifest/registry inconsistency: the top-level registry claims no required env vars, but SKILL.md metadata and code require MIMO_API_KEY (primaryEnv). Also note that using an OPENAI_API_KEY fallback could cause accidental use of an unrelated key if provided — verify which key you supply.
Persistence & Privilege
The skill does not request always:true and does not modify other skills or system-wide settings. It runs as invoked and does not request elevated system presence or unusual privileges.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install mimotts25-plus - 安装完成后，直接呼叫该 Skill 的名称或使用
/mimotts25-plus触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.2.0
mimotts25-plus v1.2.0
- 新增“用户自然语言控制指南”与规则，详细说明如何根据用户输入内容自动分配文本、风格与模型选择（自动识别情感、音色、模型类别等）。
- 指出当用户没有指定情感时可自动推断补全，强调生成内容需贴合音色描述。
- 归纳“模型选择规则”，提升“导演模式”/复杂情感的应用明确性。
- 其余内容结构和使用说明未做本质变更，重点提升了上层用户体验和自动处理流程的标准化指引。
v1.1.1
**Plus 1.1.1**
- 文档显著更新，详细补充了“核心原则”，明确 user/assistant 消息的分工与最佳实践，强调使用情感/语气指令提升自然度
- 新增/扩展自然语言控制、导演模式、音频标签及声音设计描述的具体写法与示例
- 强调自然语言指令和细节丰富性的作用，优化合成效果的实用指导
- 对现有 section 做了结构优化和内容增补，便于快速查找关键用法
- 未涉及程序代码更改，功能及接口保持与上一版本一致
v1.1.0
**Big change: Adds full compatibility with XiaomiMiMo/MiMo-Skills and new script structure.**
- Added three standalone scripts: `mimo_tts.py` (preset voices), `mimo_tts_voicedesign.py` (voice design), and `mimo_tts_voiceclone.py` (voice clone) for official CLI compatibility.
- Merged and aligned documentation with official MiMo-Skills, including detailed usage, natural language/director mode control, and voice description best practices.
- All scripts now support both `openai` library and fallback to `urllib` for API calls.
- Introduced `MIMO_API_BASE` environment variable for custom API endpoints.
- Retained and clarified the enhanced features (unified tts.py script, quick modes, extra parameters) alongside official compatibility.
v1.0.4
Plus 1.0.4 introduces a bug fix for text style tags:
- 修复 `<style>` 风格标签被 API 当成文本读出的 bug
- 风格标签自动转换为圆括号格式 `（风格）` 传给 API
v1.0.3
Refactor: clean up code structure to reduce AI scanner false positives. Move import re to top, replace sys.exit(1) with TtsError exceptions, use os.getenv().
v1.0.2
Fix: auto-parse <style> tags from inline text to prevent tags being read aloud
v1.0.1
清理示例文本，去除个人化内容，使用更规范的示例
更新版本说明，新增 Plus 版本专属功能列表
v1.0.0
升级到 v2.5-tts 模型，音质更自然
新增声音设计 (--design)：用文字描述定制专属音色
新增声音克隆 (--clone)：用音频文件复刻任意音色
新增导演模式：通过 --user-msg 实现角色/场景/指导三件套
新增 v2.5 新音色：冰糖、茉莉、苏打、白桦、Mia、Chloe
默认音色改为冰糖（可爱女声）
默认格式改为 mp3
支持 --list-models 查看所有可用模型
更新文档和参考资料
元数据
常见问题
Mimotts25 Plus 是什么？
🚀 小米 MiMo TTS 2.5 Plus — 基于 mimotts25 深度优化增强。升级到 v2.5-tts 模型，新增声音设计（文字定制音色）、声音克隆（音频复刻）、导演模式等高级功能。支持 v2.5 新音色：冰糖、茉莉、苏打、白桦等。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 128 次。
如何安装 Mimotts25 Plus？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install mimotts25-plus」即可一键安装，无需额外配置。
Mimotts25 Plus 是免费的吗？
是的，Mimotts25 Plus 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Mimotts25 Plus 支持哪些平台？
Mimotts25 Plus 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Mimotts25 Plus？
由 tear（@limingjing6666）开发并维护，当前版本 v1.2.0。
推荐 Skills