On this page
Download models, clone voices, and import audio phrases
Related page: Voices and phrases
Scenario: You want to personalize your voice experience: install the offline model and generate speech using preset voices or a newly cloned custom voice, or directly import existing sound effects and recordings as fixed phrases.
Step 1: Download and install the voice model
In , verify the model download source and click Download & install. Once the model is installed, the app can perform offline voice cloning and synthesis.
Step 2: Select or create a voice to generate speech (preset or cloned)
Go to and expand the Batch generate panel below:
- Use preset voice (default): keep Preset voice selected. The system automatically selects standard speech matching the current UI language. Simply select phrases and generate without extra setup;
- Create a new cloned voice: switch to Cloned voice, click Create new voice, enter a New voice name, and click Select reference audio (WAV) to select a clean 5–10 second voice clip (WAV format; optionally check Keep the reference audio file);
- Select the phrases you want to generate and click Start generating to batch-generate audio.
Step 3: Directly import any audio as a phrase
If you do not want to use TTS synthesis, or prefer to play existing sound effects or recordings directly:
- In the phrase management section at the top of Phrase Library, enter a Phrase name and phrase text;
- Click Import audio and select an audio file (supports wav / mp3 / flac / ogg / m4a, 0.2–30 seconds);
- The phrase status will show as Imported. Imported audio works without voice models and can be directly referenced on any feature page using Phrase library reference.