TTS Service: F5-TTS

F5-TTS is an open-source TTS service from Shanghai Jiao Tong University, supporting Chinese-English speech synthesis and voice cloning by default.

Starting from v4.04, F5-TTS is built into the software — no additional deployment needed.

In addition to the default Chinese and English, it also supports Japanese, French, German, Russian, Italian, Spanish, Hindi, and Arabic models. These community-maintained models (not official F5-TTS) can be found in the F5-TTS shared model list.

Model Download URLs

On first use, models are auto-downloaded — this may take a while and may fail. If it fails, download the .pt, .safetensors, and .txt files from the URLs below and place them in the corresponding model folder (create it if it doesn't exist).

• Official Chinese-English model (1.35G): https://huggingface.co/SWivid/F5-TTS/tree/main/F5TTS_v1_Base

Model folder: app_dir/models/models--SWivid--F5-TTS/F5TTS_v1_Base

• Japanese model (5.6G): https://huggingface.co/Jmica/F5TTS/tree/main/JA_21999120

Model folder: app_dir/models/models--Jmica--F5TTS/JA_21999120

• French model (5.4G): https://huggingface.co/RASPIAUDIO/F5-French-MixedSpeakers-reduced/tree/main

Model folder: app_dir/models/models--RASPIAUDIO--F5-French-MixedSpeakers-reduced

• German model (1.35G): https://huggingface.co/hvoss-techfak/F5-TTS-German/tree/main

Model folder: app_dir/models/models--hvoss-techfak--F5-TTS-German

• Russian model (3.4G): https://huggingface.co/hotstone228/F5-TTS-Russian/tree/main

Model folder: app_dir/models/models--hotstone228--F5-TTS-Russian

• Italian model (1.35G): https://huggingface.co/alien79/F5-TTS-italian/tree/main

Model folder: app_dir/models/models--alien79--F5-TTS-italian

• Spanish model (5.4G): https://huggingface.co/jpgallegoar/F5-Spanish/tree/main

Model folder: app_dir/models/models--jpgallegoar--F5-Spanish

• Hindi model (2.5G): https://huggingface.co/SPRINGLab/F5-Hindi-24KHz/tree/main

Model folder: app_dir/models/models--SPRINGLab/F5-Hindi-24KHz

• Arabic model (2.6G): https://huggingface.co/silma-ai/silma-tts/tree/main

Model folder: app_dir/models/models--silma-ai--silma-tts

Configuring Reference Audio

Reference audio is configured uniformly via Menu → TTS Settings → Set Reference Audio.

Steps

1. Open the "Reference Audio" settings interface.
2. In the "Reference Audio" text box, enter content in the following format:

   audio_filename.wav#text spoken in the audio file

3. Place the reference audio file in the f5-tts folder under the pyVideoTrans root directory (create it if it doesn't exist).

Example

If you have an audio file nverguo.wav containing the speech "女儿国王说话", enter:

nverguo.wav#女儿国王说话

Reference Audio Requirements

Item	Requirement
Format	WAV (recommended), MP3 also acceptable
Duration	3~10 seconds
Content	Clear pronunciation, no background noise
Text	Must match the audio content

---