Xiaomi MiMo Neural Voice & Voice Cloning Guide
Xiaomi MiMo is a next-generation neural TTS model that delivers lifelike intonation, emotional inflection, and natural pauses. It features prompt-based Voice Design and few-second Voice Cloning.
Supported Modes
| Model Mode | Capability | Description |
|---|---|---|
mimo-v2.5-tts | Preset Human Voices | High-fidelity natural voices with low latency |
mimo-v2.5-tts-voicedesign | Voice Design | Describe your desired voice using natural language (e.g. "a calm tech podcast narrator with warm cadence") |
mimo-v2.5-tts-voiceclone | Voice Clone | Upload 5-10 seconds of clear speech to replicate the target timbre |
Configuration in DuRT
- Obtain your API Key from Xiaomi MiMo Platform (api.xiaomimimo.com);
- In DuRT, go to Settings ➔ AI Capabilities ➔ Text to Speech ➔ click "Add Service" ➔ choose MiMo;
- Enter your API Key, verify Base URL (
https://api.xiaomimimo.com/v1), test connection, and save.