Brandulate OVI v1

AI-powered video + audio generation platform with a Gradio web interface.

Website: brandulate.com

Features

• Text-to-Video (T2V) generation with synchronized audio
• Image-to-Video (I2V) generation with synchronized audio
• Video extension (chain multiple segments together)
• Batch processing (generate from multiple images)
• LoRA support (up to 4 simultaneous LoRAs)
• Preset system (save and load generation settings)
• Multi-GPU support
• Memory optimization (block swap, CPU offload, FP8 quantization)
• Tiled VAE decoding for high-resolution output
• Automatic image cropping and resolution matching
• Metadata logging for all generations

System Requirements

• GPU: NVIDIA GPU with 24GB+ VRAM (48GB+ recommended)
  • Supported: RTX 3090, RTX 4090, RTX 5090, A40, A6000, L40S, RTX 6000 PRO
• CUDA: 12.8+
• cuDNN: 9.4+
• Python: 3.10
• OS: Windows 10/11, Ubuntu 20.04+
• Disk: 100GB+ for models

Quick Start

Windows

1. Run Windows_Install_and_Update.bat
2. Run Windows_Start_App.bat
3. Open the Gradio URL shown in terminal

RunPod / Massed Compute

See Runpod_Instructions_READ.txt or Massed_Compute_Instructions_READ.txt.

Project Structure

OVI/
  premium.py          # Main application (Gradio UI + generation logic)
  ovi/                # Core engine module
    ovi_fusion_engine.py  # Fusion engine for video+audio generation
    configs/          # Model and inference configs
    utils/            # Utility modules (LoRA, I/O, model loading)
  ckpts/              # Model checkpoints (downloaded separately)
  outputs/            # Generated videos
  loras/              # LoRA weight files
  presets/            # Saved generation presets

Models

Models are hosted at huggingface.co/GWD99/ovi-dependencies and downloaded automatically during installation.

License

Apache 2.0