Skip to content


At "Projects" page you can view all projects and datasets you have.


We have two-level structure for simplicity: projects and datasets inside project. Dataset contains images and their annotations. One image logically can not be assigned to few datasets or projects.

Images are not duplicated.

You can upload images multiple times, only one version will be stored inside system. This approach allows us effectively work with cloud storage. Also this has a good effect on system performance and backup speed.

Classes and tags are assigned to the project and available for all datasets inside it.

Example of structure. This user has 3 projects, three datasets inside project1, one dataset inside project2, and two datasets inside projectN.

├── project1
│   ├── dataset1
│   ├── dataset2
│   └── datasetK
├── project2
│   └── dataset777
└── projectN
    ├── dataset_999
    └── ds333

List of Projects

Here is the example of "Projects" page.

Project page

Just click to the project to view its datasets.

How to go to annotation tool?

If you click to dataset, you'll be redirected to Annotation tool.

As you can see from the screenshot above, user can view different information about project:

  • Datasets tab — list of all datasets in the project

  • Info tab — meta information of project

  • Statistics — extended statistics about images, objects, their various distributions.

  • Classes — list of all available object classes in the project.

  • Tags

    • Image tags — list of all available image tags in the project.

    • Object tags — list of all available object tags in the project.

  • Setting — json-configurable project settings. This settings allows to configure behaviour for most of the tools in "Annotation" module