Unprofessional Audio & Video Processing Suite
English | 繁體中文
A free, offline, and ready-to-use AI audio & video workstation developed by a single person.
| Feature | Description | Core Technology |
|---|---|---|
| 🎬 Video/Audio Downloader | One-click download of videos from 1000+ platforms, supports lossless 4K | yt-dlp + FFmpeg |
| 🎤 Vocal & Accompaniment Separation | AI separates vocals, drums, bass, and accompaniment | Meta Demucs + MDX-Net |
| 🎵 KTV Recording & Synthesis | Load accompaniment for real-time singing, automatic mixing | WASAPI + AI Noise Reduction |
| 📝 AI Subtitle Generation | Auto transcription + translation + bilingual subtitles + hardburning | OpenAI Whisper + Pyannote |
| 🗣️ Microsoft TTS | Free AI neural network voice synthesis, multi-language, multi-character | Microsoft Edge-TTS |
| 🧬 Voice Cloning | Mimic voice tone and intonation with just 3~10 seconds of audio sample | GPT-SoVITS |
| 🔄 RVC Voice Changer Inference | AI cover generation, replacing vocal timbre while preserving emotion | RVC |
| 🎙️ RVC Real-time Voice Changer | Real-time voice changing from microphone, supports VB-Cable routing | RVC Real-time |
| 🏋️ RVC Model Training | Free training of your own custom voice models via Google Colab | Applio RVC |
| Tool | Description |
|---|---|
| 🎙️ Recording Assistant | Lightweight recording + real-time waveform display + AI noise reduction |
| 🔄 Universal Format Converter | High-speed lossless conversion between video/audio formats |
| 🎵 Audio Extractor | Extract high-quality stereo audio tracks from video files |
| ✂️ Lossless Cutter | Quick cutting without re-encoding, zero loss in quality |
| 🔗 Lossless Merger | One-click merge of video and audio tracks, supports mixing |
| 📦 Intelligent Video Compressor | FFmpeg CRF encoding, compresses file size with no perceptual visual loss |
| 🤫 Auto Silence Cut | AI detects silent segments and clips out pauses automatically |
| Item | Recommended Spec |
|---|---|
| Operating System | Windows 10/11 (64-bit) |
| Graphics Card | NVIDIA RTX Series (Highly Recommended) |
| RAM | 8GB or more |
| Disk Space | Full Version: ~38GB / Medium Version: ~25GB |
⚠️ Currently only Windows is officially supported. You can run it without an NVIDIA GPU (it automatically falls back to CPU mode), but the performance will be slower.
| Item | Full Version (~38GB) | Medium Version (~25GB) |
|---|---|---|
| Voice Cloning (GPT-SoVITS) | ✅ | ❌ |
| All Other Features | ✅ | ✅ |
| Upgrade Path | Simply replace Studio0808.exe |
Simply replace Studio0808.exe |
- Full Version (Includes Voice Cloning): Google Drive Download
- Full Version Backup Link: Google Drive Backup
- Medium Version (No Voice Cloning): Google Drive Download
After downloading, extract the ZIP file and run Studio0808.exe directly. No installation required.
Complete feature explanation and usage guide: https://begin0808.github.io/studio0808_video/
Bilingual version (Traditional Chinese & English) is available.
| Technology | Purpose |
|---|---|
| PyTorch v2.6.0+cu124 | Core AI computation engine (RTX 50 series ready) |
| FFmpeg | Video/audio encoding, decoding, merging, and cutting |
| Demucs (htdemucs) | Meta vocal separation model |
| faster-whisper | AI speech-to-text transcription and subtitle generation |
| Pyannote | Speaker diarization |
| GPT-SoVITS | Voice cloning and text-to-speech synthesis |
| RVC | Retrieval-based Voice Conversion (voice changing) |
| Edge-TTS | Microsoft cloud-based speech synthesis |
| yt-dlp | Video and audio streaming downloader |
| Torchcrepe | High-precision pitch estimation |
| CustomTkinter | Modern GUI framework |
This GitHub repository contains the complete Python source code. To keep the repository lightweight, large model weights and external binary tools are not included (excluded in .gitignore).
If you want to run from source or package the program on Windows, please follow these steps to download the necessary binaries:
- Download the latest Full Version ZIP from the download section above and extract it.
- Copy the following items from the extracted folder directly into your cloned repository folder:
models/(contains all AI model weights)tools/(contains executables like ffmpeg, yt-dlp, etc.)modules/hubert_base.ptandmodules/rmvpe.pt(RVC core models)GPT-SoVITS/(the complete GPT-SoVITS environment)
- Copy
modules/configs/config.json.exampleand rename it toconfig.json. - Install dependencies:
pip install -r requirements.txt - Run the main application:
python Studio0808_Video.py
If you are a macOS user with Python development experience, this project supports porting, running, and packaging on macOS. Please refer to the standalone guide for detailed environment configuration and instructions: 👉 macOS Development, Testing & Packaging Guide (README_macOS.md)
- Discord: Join the Community
- For questions, bug reports, or feature requests, feel free to start a conversation in our Discord!
Studio0808 is a completely free suite with no trials, no feature limitations, and no ads.
If you find it helpful, feel free to buy the developer a coffee:
- LINE ID:
begin0808(Jia-En Li) - Supports LINE Pay and TWQR transfers
- This software is for personal learning, research, and academic exchange only.
- Commercial copyright infringement, voice cloning for fraud, or spreading misinformation is strictly prohibited.
- Users assume full legal responsibility for their own actions.
The documentation website (docs) is hosted via GitHub Pages.
The software application itself is released independently as a free desktop application.
© 2026 Studio0808 Team. All rights reserved.
Made with ❤️ in Taiwan