Skip to content

This is a webui and api project for the kokoro TTS project, supporting voiceovers in 8 languages: Chinese, English, Japanese, French, Italian, Portuguese, Spanish, and Hindi.

Project address https://github.com/jianchang512/kokoro-uiapi

Web Interface

The default UI address after startup: http://127.0.0.1:5066

  • Supports voiceovers for text and SRT subtitles
  • Supports online listening and downloading
  • Supports subtitle alignment

Installation Method

Windows

For Win10/11, you can directly download the integrated package and double-click start.bat to start. If you need GPU acceleration, please ensure you have an NVIDIA graphics card and CUDA12 installed.

Baidu Netdisk Download Address: https://pan.baidu.com/s/1jTB84E3-gaLqFrl32f4sDw?pwd=xnwp

GitHub Download (excluding models, requires online download via VPN): https://github.com/jianchang512/kokoro-uiapi/releases/download/v0.1/kokoro-uiapi-noModels-v0.2.7z

Linux/MacOS

First, ensure that python3.8+ is installed on the system, it is recommended to use 3.10-3.11

On Linux, use apt install ffmpeg or yum install ffmpeg to pre-install ffmpeg

On MacOS, use brew install ffmpeg to install ffmpeg

  1. Pull the source code git clone https://github.com/jianchang512/kokoro-uiapi
  2. Create and activate a virtual environment
    cd kokoro-uiapi
    python3 -m venv venv
    . venv/bin/activate
  3. Install dependencies pip3 install -r requirements.txt
  4. Start python3 app.py

Usage in pyVideoTrans

  1. First, start this project. For the Windows integrated package, double-click start.bat. For source code installation, execute python3 app.py.

  2. Upgrade pyVideoTrans to v3.48+, open Menu -- TTS Settings - Kokoro TTS -- fill in http://127.0.0.1:5066 for the HTTP address.

OpenAI API Compatibility

The API is compatible with OpenAI TTS.

The default API address after startup: http://127.0.0.1:5066/v1/audio/speech

Request method: POST Request data: application/json

{
		input: text to be voiced,
		voice: voice role,
		speed: speech rate, default is 1.0
}

Successful return of mp3 audio data

OpenAI SDK Usage Example

from openai import OpenAI
client = OpenAI(
    api_key='123456',
    base_url='http://127.0.0.1:5066/v1'
)

try:
    response = client.audio.speech.create(
		model='tts-1',
        input='Hello, dear friends',
        voice='zf_xiaobei',
        response_format='mp3',
        speed=1.0
	)
    with open('./test_openai.mp3', 'wb') as f:
        f.write(response.content)
    print("MP3 file saved successfully to test_openai.mp3")
except Exception as e:
    print(f"An error occurred: {e}")

Role List

English voice roles:


af_alloy
af_aoede
af_bella
af_jessica
af_kore
af_nicole
af_nova
af_river
af_sarah
af_sky
am_adam
am_echo
am_eric
am_fenrir
am_liam
am_michael
am_onyx
am_puck
am_santa
bf_alice
bf_emma
bf_isabella
bf_lily
bm_daniel
bm_fable
bm_george
bm_lewis

Chinese roles:

zf_xiaobei
zf_xiaoni
zf_xiaoxiao
zf_xiaoyi
zm_yunjian
zm_yunxi
zm_yunxia
zm_yunyang

Japanese roles:

jf_alpha
jf_gongitsune
jf_nezumi
jf_tebukuro
jm_kumo

French roles: ff_siwis

Italian roles: if_sara,im_nicola

Hindi roles: hf_alpha,hf_beta,hm_omega,hm_psi

Spanish roles: ef_dora,em_alex,em_santa

Portuguese roles: pf_dora,pm_alex,pm_santa

Proxy VPN

Source code deployment requires downloading the voice pt file from huggingface.co. You need to set up a global or system proxy in advance to ensure accessibility.

You can also download the model in advance and extract it to the directory where app.py is located.

Model download address https://github.com/jianchang512/kokoro-uiapi/releases/download/v0.1/moxing--jieya--dao--app.py--mulu.7z

Credit