Woyage AI Avatar Backend

This FastAPI service accepts an MP3 file, uses a lip-sync model to generate a speaking avatar video, uploads it to S3, and returns the path.

Installation

Step 1: Open PowerShell or CMD and clone the repo

git clone https://github.com/adprimak/woyage-generate-avatar.git
cd woyage-generate-avatar

Step 2: Clone the fantasy-talking repo

git clone https://github.com/Fantasy-AMAP/fantasy-talking

Step 3: Download models to woyage-generate-avatar/models/ as documented here

Step 4: Create a virtual environment

cd /path/to/woyage-generate-avatar
python -m venv venv

Step 5: Activate the virtual environment

venv\Scripts\activate

Step 6: Install dependencies

pip install --upgrade pip
pip install -r requirements.txt
pip install -r fantasy-talking/requirements.txt

Step 7: Add AWS credentials to .env.

BUCKET_NAME=your-bucket-name
AWS_ACCESS_KEY_ID=your-access-key-id
AWS_SECRET_ACCESS_KEY=your-secret-access-key
AWS_DEFAULT_REGION=us-east-1

Step 8: Ensure the avatar image is placed in static/avatar.png.

Run

Step 1: Start the FastAPI server

uvicorn app.main:app --reload

Step 2: Open a browser and go to:

http://localhost:8000/docs

Step 3: Expand the POST /video/generate_avatar endpoint

Step 4: Click "Try it Out"

Step 5: Upload an mp3 file

Step 6: Click "Execute"

Folder Structure

woyage-generate-avatar/
├── app/
│   └── main.py
├── static/
│   └── avatar.png
├── fantasy-talking/
│   └── models/
│   └── requirements.txt
├── requirements.txt
└── README.md