Models run
in your browser.
Use WebGPU for similar-object search on compatible devices, and run supported local detection and classification models in your browser.
Explore local AIUse the GPU
you already have.
Detect and classify with local AI, review with your team, then export or train.
No credit card required.
Label
Start with local or cloud AI. Refine every annotation.
Review
Assign images and bring your team into the review.
Export or train
Export your dataset, or train a model with your own GPU account.
Your workflow.
In one workspace.
Create segmentation annotations with clicks, or select an object and find similar ones in the same image. WebGPU accelerates these workflows on compatible devices.
Segmentation / 1 of 2
Models run
More images.
Same workflow.
Process image batches with managed Cloud AI detection or classification, and follow progress as annotations are saved to your project.
See the included Cloud AI limitsYour images. Your field.
Start free. Make room for more.
Free Tier
For your first datasets and small team projects.
$0forever
Start freeIncluded in Free Tier
- 3 team members · 2 projects
- 2,500 images · 5 GB managed storage total
- Managed cloud AI: 150 runs per day
- Unlimited local AI-assisted labeling
- Unlimited review workflows
- All supported export formats
- 3 version snapshots per project
Coming soon
Pro
For larger datasets, more cloud AI, and a growing team.
$49/month + tax
Included in Pro
- 10 team members · 10 projects
- 50,000 images · 100 GB managed storage total
- Managed cloud AI: 2,000 runs per day
- Unlimited local AI-assisted labeling
- Unlimited review workflows
- All supported export formats
- 20 version snapshots per project
A few things to know.
Which AI models can I use for labeling?
Local zero-shot detection and classification models that run in your browser, or managed Cloud AI detection and classification for larger batches.
How do I assign work to my team?
Create an assignment for a single image or a range, add instructions and a due date, and track it until the work is reviewed.
How does review work?
Invite teammates as owners, reviewers, or annotators, assign images, and approve annotations or return them with a note.
Can I train a model after labeling?
Yes. Connect your own GPU provider account to compare offers and run training jobs, and the provider bills compute directly to you.
Can I track changes to my dataset?
Save named versions of a project and compare two of them to see exactly which annotations were added, changed, or removed.
What export formats are supported?
COCO JSON, YOLO, Pascal VOC XML, LabelMe JSON, CVAT XML, JSONL, CSV, and TSV, with annotation metadata, confidence scores, and creator information preserved.
Can I use LabelOp without an account?
Yes. The format converter, annotation merger, dataset splitter, health report, and label mapper are free and need no login.
Do I need a GPU?
No. Local models use your device when WebGPU is available.
