LabelU
opendatalab.github.io
- Free plan
- Windows
- macOS
- Web
- Self-hosted
What LabelU is
LabelU is an open-source platform for annotating image, video, and audio data. Image tools include 2D bounding boxes, semantic segmentation, polylines, and keypoints; video and audio tools support segmentation, classification, and information extraction. Users can load pre-annotated data and refine or adjust it. Image AI auto-annotation can detect and segment objects, run batch tasks, and report progress in real time. Reference servers are provided for Florence-2, GroundingDINO with SAM ViT-B, and SAM 3. The listed export formats include JSON, COCO, and MASK, and files can be imported from S3-compatible storage such as AWS S3 and MinIO. Users can access an online instance and browser-based toolkit or install LabelU locally with pip. Local setup uses Python 3.11, runs at localhost:8000, and defaults to SQLite, with optional MySQL support. The reference models have stated resource needs: about 4 GB of memory for Florence-2 and GroundingDINO with SAM ViT-B, and about 8 GB for SAM 3; SAM 3 requires CUDA 12.6 or later. Published by the OpenDataLab GitHub organization under the Apache 2.0 license, LabelU is free. The README invites users to join the OpenDataLab official WeChat group.
Who it suits
LabelU suits teams annotating image, video, or audio data, especially when they want model-assisted image labeling or local deployment. Users considering the reference AI models should check their stated memory and CUDA requirements.
Good to know, and what to watch
- Supports image, video, and audio annotation.
- Image tools include boxes, segmentation, polylines, and keypoints.
- Imports from S3-compatible storage.
- Exports include JSON, COCO, and MASK.
- Local setup specifies Python 3.11.
- SAM 3 requires about 8 GB memory and CUDA 12.6 or later.
The verdict
LabelU covers multiple media types and offers image auto-annotation alongside manual and pre-annotated workflows. Its local setup and AI model requirements are worth checking before choosing a deployment approach.
What the maker publishes
- AI annotation
- Image AI auto-annotation supports batch tasks and real-time progress, with reference servers for Florence-2, GroundingDINO plus SAM ViT-B, and SAM 3. source
- AI auto-annotation
- AI services can automatically detect and segment image objects, with batch annotation and real-time progress tracking. source
- AI model options
- The project provides reference servers for Florence-2, GroundingDINO with SAM ViT-B, and SAM 3. source
- AI resource limits
- The reference AI models list minimum memory requirements of about 4 GB for Florence-2 and GroundingDINO with SAM ViT-B, and about 8 GB for SAM 3; SAM 3 requires CUDA 12.6 or later. source
- Audio annotation
- Audio tools support segmentation, classification, and information extraction. source
- Database option
- The default database is SQLite, and the maker documents optional MySQL support. source
- Database option
- SQLite is the documented default database, and MySQL support is available as an optional installation. source
- Deployment
- LabelU can be installed locally with pip and run at localhost:8000; the README specifies Python 3.11. source
- Export formats
- The maker lists JSON, COCO, and MASK among supported export formats. source
- Export formats
- The project says it supports exporting annotation data in JSON, COCO, and MASK formats. source
- Image annotation
- Image tools include 2D bounding boxes, semantic segmentation, polylines, and keypoints. source
- Image tools
- Image annotation includes 2D bounding boxes, semantic segmentation, polylines, and keypoints. source
- License
- The project is released under the Apache 2.0 license. source
- License
- The repository states that LabelU is released under the Apache 2.0 license. source
- Maker
- The LabelU repository is published under the OpenDataLab GitHub organization. source
- Online access
- The maker links to an online LabelU instance and a browser based annotation toolkit. source
- Pre-annotations
- Users can load pre-annotated data and refine or adjust it. source
- Purpose
- LabelU is a multimodal data annotation platform for image, video, and audio tasks. source
- Purpose
- LabelU is a multimodal data annotation platform for image, video, and audio tasks. source
- Self-hosting
- The README gives local installation steps using Python 3.11, pip, and a local server at localhost:8000, with Windows Anaconda Prompt instructions and a macOS Intel note. source
- Storage integration
- LabelU can import files from S3-compatible storage, including AWS S3 and MinIO. source
- Storage integration
- LabelU can import data from S3-compatible object storage, including AWS S3 and MinIO. source
- Support channel
- The README invites users to join the OpenDataLab official WeChat group. source
- Video and audio
- The maker describes video and audio annotation for segmentation, classification, and information extraction. source
Checked against the maker's pages on 9 Oct 2026.