Private · runs on your device
Annotate with on-device AI, 100% on your machine.
Segment with one click, find objects by text, and accelerate annotation with on-device models. Your media, labels and annotations stay on your device.
- No account
- No cloud upload
- Runs locally
- Web and native builds
SAM 2.1 · on this device
Instance segmentation
Local-first
Your data has nowhere to go — and that is the point
AnnotateIt is a local-first application. Projects, media, labels and annotations are stored on your device, and AI inference runs on your hardware.
Images and videos you import stay in local storage on your device.
Manual tools and on-device AI models annotate locally. No account, no cloud backend.
You export standard dataset formats to wherever you choose.
- Projects, media, labels and annotations are stored locally on your device.
- AI-assisted annotation runs on your device — CPU/WASM everywhere, WebGPU where available.
- The web app downloads application and model files from the network; your datasets are never uploaded.
- There is no cloud sync, no accounts and no server-side collaboration — by design, not as a missing feature.
Annotation tasks
Built for computer-vision datasets
Create a project for the task you are training for — every project type has purpose-built tools and exports.
Object detection
Draw and refine bounding boxes, or let a detection model propose them.
Instance segmentation
Polygons and masks — drawn by hand or with one click of Segment Anything.
Keypoint detection
Skeleton templates with joints and edges, with pose assistance for people.
Classification
Single-label, multi-label and hierarchical projects with fast keyboard tagging.
Images and video frames
Annotate photos or extract and label frames from video, in the same project.
On-device AI
Smart tools, honest about where they run
Every model below executes on your machine. Built-in engines are ready immediately; larger ones download on demand — and still run locally.
One-click segmentation
Click an object to get a mask, refine with clicks, or turn a box into a mask. Segment Anything is built in; SAM 2.1 Tiny, Small and Large are optional upgrades.
- Segment AnythingBuilt-in
- SAM 2.1On-demand download
Text-prompt detection
Type what you are looking for — open-vocabulary detection turns your words into boxes across the image.
- CLIPBuilt-in
- Grounding DINO TinyOn-demand download
Zero-shot classification & search
Classify without training and search your media semantically, in multiple languages.
- SigLIP 2On-demand download
Pose assistance
Draw a person box and get 17 COCO keypoints placed onto your skeleton template.
- RTMPose-mOn-demand download
Detection assistant
Find more of what you already labeled — visual prompting proposes similar objects for review.
- Built-in engineBuilt-in
Your own ONNX models
Import your own model through a guided wizard — inspect, map labels, test, save. Supports YOLOv8 detection, segmentation and pose, RT-DETR, DETR and image classifiers; other output shapes are rejected. Experimental.
- Custom ONNXYour file
Engine availability depends on platform and hardware: everything runs on CPU/WASM, larger models prefer WebGPU, and mobile builds exclude the heaviest engines.
Real product
The actual interface, not a mockup
Every screenshot below is the shipping application.
Workflow
From raw media to training-ready dataset
No pipeline setup, no upload step — the whole loop happens on your machine.
Create a project
Pick a task: detection, segmentation, keypoints or classification.
Import media
Drop in images or a video — frames are extracted locally.
Annotate
Label manually or let on-device AI do the heavy lifting.
Export
Download a standard dataset format, ready for training.
Interoperability
Standard formats in, standard formats out
Move datasets between AnnotateIt and your training stack without converters.
Availability depends on project type and dataset contents — Supervisely Video, for example, applies when a dataset contains video.
Format details in Docs →- COCO
- YOLO
- Pascal VOC
- Datumaro
- Supervisely Video
- Plain ZIP
Platforms
Same product, your choice of surface
Start in the browser in seconds, or install a native build for offline work and store-managed updates.
Web
Runs in a modern browser. Models download on demand and are cached locally.
Open web appWindows
Native desktop build with bundled AI engines, from the Microsoft Store.
Microsoft StoremacOS
Native build for Apple silicon Macs, from the Mac App Store.
Mac App StoreiPhone & iPad
Annotate on the go with touch-first tools, from the App Store.
App Store
Frequently asked questions
Does my data leave my device?
No. Datasets, media, labels and annotations stay on your machine, and AI inference for smart tools runs locally. The web app fetches application and model files from the network, but never uploads your content.
Do I need an account?
No. There is nothing to sign up for — open the app and create a project.
Does the web version work offline?
Once the app and the models you use are cached, annotation work runs locally. The native desktop builds are the best choice for fully offline sessions.
What is the difference between web and native?
The same product. Native builds bundle the default AI engines, work offline and get store-managed updates; the web version downloads engines on demand and is the fastest way to try AnnotateIt.
Which annotation tasks are supported?
Object detection, instance segmentation, keypoint detection, and single-label, multi-label and hierarchical classification — for images and video frames.
Which formats can I import and export?
COCO, YOLO, Pascal VOC, Datumaro, Supervisely Video and Plain ZIP (media only), depending on project type.
Can I use my own ONNX model?
Yes — an experimental five-step wizard, opened inside a project, imports YOLOv8 detection, segmentation and pose models, RT-DETR, DETR and plain image classifiers (up to 512 MB). It inspects the file, maps its classes to your labels and requires a successful test run before saving. Other output shapes are rejected — it is not "any ONNX".
Does AnnotateIt train models?
No. AnnotateIt is for building datasets: annotate locally, then export and train in the stack of your choice.
Is it free?
You can use AnnotateIt with up to 5 active projects for free. Pro is available in the Microsoft Store and Mac App Store builds.
Open AnnotateIt in your browser
No sign-up, no install, no upload — create your first project right in the browser.