System Requirements
Vision AI Label Studio is a local-first desktop application. The core labeling experience is lightweight and runs comfortably on most modern machines. The optional AI features (automatic detection, segmentation, and the offline copilot) benefit greatly from a GPU but also work on CPU.
Operating system
| Platform | Supported versions |
|---|---|
| Windows | Windows 10 or Windows 11 (64-bit) |
| macOS | macOS 11 Big Sur or later (Apple Silicon and Intel) |
See the desktop installation guide for how to get past the unsigned-app warning on each platform.
Hardware
| Component | Minimum | Recommended |
|---|---|---|
| CPU | 64-bit dual-core | Quad-core or better |
| RAM | 4 GB | 8 GB+ (16 GB for large datasets or AI) |
| Disk | ~500 MB for the app | Extra space for your images and projects |
| Display | 1280 × 720 | 1920 × 1080 or higher |
Your projects, images, and annotations are stored locally, so plan disk space around the size of your image datasets.
Optional: GPU for AI acceleration
The AI features run through ONNX Runtime. They will run on CPU out of the box, but an NVIDIA GPU makes detection and segmentation dramatically faster.
| Requirement | Details |
|---|---|
| GPU | NVIDIA GPU with up-to-date drivers |
| CUDA | CUDA 12 runtime |
| cuDNN | cuDNN 9 (for CUDA 12) |
| Extra disk | ~1 GB for the downloaded GPU runtime + cuDNN |
The app can download and install the GPU runtime for you (Windows) — see the AI & GPU setup guide for the full walkthrough and troubleshooting. To learn what the AI features do, see the AI Copilot guide.
No GPU? That's fine — AI detection still works on CPU, just slower. You can add a GPU runtime later at any time.