AI Livestreamer with VTube Studio

An AI livestreamer that generates TTS content and animates a Live2D avatar using VTube Studio.

Setup

Install dependencies:

pip install -r requirements.txt

Create .env file:

OPENAI_API_KEY=your_key_here

TTS Configuration - Easy Switching!

Open main.py and change these lines (around line 27):

# Use TikTok TTS (free, no setup needed):
TTS_PROVIDER = "tiktok"
VOICE_OPTIONS = {"tiktok": ['en_us_001']}

# OR use ElevenLabs (better quality, needs API key):
# TTS_PROVIDER = "elevenlabs"
# VOICE_OPTIONS = {"elevenlabs": ['rachel']}

To use ElevenLabs:

1. pip install elevenlabs
2. Get API key from https://elevenlabs.io
3. Add ELEVENLABS_API_KEY=your_key to .env
4. Change config in main.py to use "elevenlabs"

Configuration

Edit settings in main.py:

# TTS - Change provider here
TTS_PROVIDER = "tiktok"  # or "elevenlabs"

# Avatar
ENABLE_AVATAR = True
AVATAR_GAIN = 2.5
AVATAR_SMOOTHING = 0.7

# Text Display
ENABLE_TEXT_DISPLAY = True
TEXT_DISPLAY_STYLE = "progressive"

OBS Text Display Setup

1. Add Source → Text (GDI+)
2. Check "Read from file"
3. Select current_text.txt
4. Style and position it

Usage

python main.py

Troubleshooting

Avatar doesn't move: Increase AVATAR_GAIN to 3.0-4.0

Movement is jittery: Increase AVATAR_SMOOTHING to 0.8-0.9

TikTok TTS fails: Proxy services may be down, try again later

ElevenLabs fails: Check API key in .env and account credits