RabiSpeech IndexTTS2
SkillMediaLets your agent generate Chinese speech and cloned voices locally using the IndexTTS2 text-to-speech model.
Available today. Use it from your connected AI after setup.
No other account needed.
Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.
Then ask your AI: use the RabiSpeech IndexTTS2 skill
About this skill
Generates Chinese speech or voice-cloned audio via the local IndexTTS2 model through RabiSpeech. Use when the user specifies IndexTTS2, a Chinese reference audio, emotion/style instructions, a persona voice, or local WAV verification. Do not directly run the skill's worker, start OumuQ, or call clou
What this skill tells your AI
The instructions your AI receives, as published by vb2250158/rabiroute in skills/indextts2-audio/SKILL.md and read by ahel’s review.
IndexTTS2 由 RabiSpeech 管理，模型 id 为 local-tts/indextts2。人格参考音频位于 RabiRoute/data/roles/<RoleId>/voice/，生成结果不得写回参考库。
流程
先读 TTS 路由。speechBaseUrl 必须从当前服务配置取得；下例中的 $speechBaseUrl 由该发现步骤赋值。
- 检查
GET <speechBaseUrl>/v1/models/local-tts/indextts2。 - 传入
voice=<RoleId>；RabiSpeech 从人格 voice index 选择并缓存参考音频。 - 需要情绪或语气时用
instructions描述，保持简短、可执行。 - 验证用
play=false并保存返回 WAV；对话用play=true、session_id，进入全局 FIFO。
$body = @{
model = 'local-tts/indextts2'
input = '你好，这是 IndexTTS2 本地语音。'
voice = '<RoleId>'
response_format = 'wav'
speed = 1.0
language = 'zh'
instructions = '自然、温和、清楚。'
play = $false
} | ConvertTo-Json -Compress
Invoke-WebRequest -Method Post -Uri ($speechBaseUrl.TrimEnd('/') + '/v1/audio/speech') -ContentType 'application/json' -Body $body -OutFile '.\indextts2.wav'
不要从技能目录运行旧 indextts2_worker.py；模型环境、CUDA、参考音频组合和输出缓存统一属于 RabiSpeech。
Signals
- GitHub stars
- 502
- Last commit
- Sep 2026
Advanced
- Item type
- skill
- Key
indextts2-audio- Source
- github.com/vb2250158/rabiroute