- Home
- AI Image Annotation Tools
- YoloLabel
YoloLabel
What it is
YoloLabel is a free desktop GUI for marking object bounding boxes in images used to train YOLO neural networks. It loads JPG or PNG files from a directory and creates boxes through a two-click method. Annotation tools let users move, resize, copy, paste, undo, and redo labels, which can be exported as YOLO TXT. For assisted labeling, YoloLabel runs local inference with Ultralytics detection models exported to ONNX, including YOLOv5, YOLOv8, YOLO11, YOLO12, and YOLOv26. A loaded model can label the current image or process a dataset in batches. The application also connects to YoloLabel AI for cloud open-vocabulary detection through an API key and optional prompt; cloud requests handle up to 20 images per batch. Prebuilt downloads cover Windows x64, Linux x64, and macOS Apple Silicon. The desktop repository uses the MIT License. Building from source requires Qt 6, while ONNX Runtime is optional unless local auto-labeling is needed. The README warns that moving the horizontal image slider does not automatically save the last processed image.
For and against
For
- Free desktop annotation tool
- Supports manual box editing and history
- Runs local ONNX model inference
- Batch-labels datasets with a loaded model
- Prebuilt downloads cover three platforms
Against
- Slider movement may not save the last image
- Source builds need Qt 6
- Local auto-labeling requires ONNX Runtime
- Cloud free tier allows 100 images monthly
The verdict
YoloLabel provides manual bounding-box annotation, local model assistance, optional cloud detection, and YOLO TXT export across major desktop platforms. Its source-build requirements and cloud quota are the main constraints to consider.
Who it suits
It suits developers and computer-vision annotators preparing YOLO training data on Windows, Linux, or Apple Silicon Macs. Users wanting local ONNX labeling or optional cloud detection can choose between those workflows.
On the record
- Annotation
- It supports manual bounding box labeling and uses a two left-click method to create boxes.github.com · read 30 Sep 2026
- Batch labeling
- With a loaded ONNX model, users can auto-label the current image or batch-process all images in the dataset.github.com · read 30 Sep 2026
- Build requirement
- Building from source with auto-label support requires ONNX Runtime; without it, the app works without that feature.github.com · read 30 Sep 2026
- Cloud API
- YoloLabel AI provides a REST API that accepts images and prompts and returns detections and YOLO-format labels.yololabel.com · read 30 Sep 2026
- Cloud batch limit
- Cloud Auto Label All submits images in batches of up to 20 per request.github.com · read 30 Sep 2026
- Cloud data retention
- The cloud service says it retains job metadata for 90 days and account and usage records while an account is active.yololabel.com · read 30 Sep 2026
- Cloud image handling
- The cloud service privacy policy says uploaded images are processed in memory, discarded after inference, and not used to train models.yololabel.com · read 30 Sep 2026
- Cloud integration
- YoloLabel integrates with yololabel.com for cloud open-vocabulary object detection, using an API key and optional detection prompt.github.com · read 30 Sep 2026
- Cloud limits
- The cloud service terms state that the free tier includes 100 images per month with no SLA, unused quota does not roll over, and over-limit requests return HTTP 402.yololabel.com · read 30 Sep 2026
- Cloud security
- The cloud privacy policy says it uses HTTPS, bcrypt password hashing, and short-lived JWTs with refresh token rotation.yololabel.com · read 30 Sep 2026
- Download platforms
- The README lists prebuilt downloads for Windows x64, Linux x64, and macOS on Apple Silicon.github.com · read 30 Sep 2026
- Downloads
- Prebuilt desktop downloads are listed for Windows x64, Linux x64, and macOS Apple Silicon.github.com · read 30 Sep 2026
- Image formats
- The README says to load .jpg or .png images from a directory.github.com · read 30 Sep 2026
- Image tools
- The app includes real-time contrast adjustment and a usage timer that runs while its window is focused.github.com · read 30 Sep 2026
- License
- The desktop repository is licensed under the MIT License, which permits use, modification, distribution, and sale subject to its stated conditions.github.com · read 30 Sep 2026
- Local auto-labeling
- It can run local inference with Ultralytics detection models exported to ONNX, including YOLOv5, YOLOv8, YOLO11, YOLO12, and YOLOv26.github.com · read 30 Sep 2026
- Maker
- The maker’s GitHub profile identifies developer0hye as Yonghye Kwon.github.com · read 30 Sep 2026
- Manual annotation
- It uses a two-click method to create boxes and includes tools to move, resize, copy, paste, undo, and redo annotations.github.com · read 30 Sep 2026
