Private, local-first computer vision annotation with on-device AI.
AnnotateIt is a computer vision annotation tool that runs entirely on your machine. Projects, media, labels, and annotations stay on your device, and AI-assisted tools run inference locally — your datasets are never uploaded to a server.
Open the web app · Try a demo project · Download native apps · Read the documentation · Community and support
AnnotateIt supports annotation of images and video frames for the following tasks:
- Object detection
- Instance segmentation
- Keypoint detection
- Single-label classification
- Multi-label classification
- Hierarchical classification
AI-assisted annotation runs locally on your device. Built-in tools include:
- Segment Anything and SAM 2.1 — one-click and promptable segmentation
- Grounding DINO — text-prompted object detection
- SigLIP 2 — zero-shot classification and semantic search
- RTMPose — pose assistance for keypoint annotation
Importing custom ONNX models is available as an experimental feature and supports the compatible architectures described in the documentation.
The web app downloads application and model files from the network, but it does not upload your datasets. Your images, video, and annotations remain on your device.
Depending on the project type, AnnotateIt can import and export datasets in:
- COCO
- YOLO
- Pascal VOC
- Datumaro
- Supervisely Video
- Plain ZIP
| Platform | Get it |
|---|---|
| Web browser | app.annotateit.ai |
| Windows | Microsoft Store |
| macOS | Mac App Store |
| iPhone and iPad | App Store |
annotateit-community is the public space for everything that needs a public trail:
- Report a bug — reproducible problems, with the version, platform and steps that make them fixable
- Request a feature — describe the workflow you need
- Import or export problems — COCO, YOLO, Pascal VOC, Datumaro, Supervisely Video and Plain ZIP compatibility
- Release notes and the public roadmap
Please keep private datasets, user images and personal information out of public issues. Security vulnerabilities go to umno.annotateit@gmail.com instead — never a public issue.
AnnotateIt is proprietary software. The application source code is not published in this organization's repositories; the public repositories here contain organization-level documentation and policies only.
- Product support: umno.annotateit@gmail.com
- Security reports: umno.annotateit@gmail.com (mark the subject as a security report) — see the security policy