How to integrate CVAT interface into my own system

[imane-pro]

I am working on a project in the automotive insurance field (vehicle damage expertise).
We need to annotate car damage images, and I would like to know if it is possible to integrate the CVAT annotation interface directly into our own web system (instead of using CVAT as a separate platform).
Specifically:

Can CVAT be embedded in an iframe or integrated via API?

Is there any documentation or recommended approach for linking CVAT tasks with a custom web application?

Would you suggest using CVAT directly or building a custom annotation frontend with the CVAT backend?

Thank you very much for your help!

[bsekachev]

Hello

It can embedded via iframe, however this process is not documented.
You may also use REST API and wrap it into your own frontend. Specification may be found on <INSTANCE_URL>/api/swagger.
Alternatively, you may use some our components like cvat-core, cvat-canvas which implement basic business logic of working with objects (server interaction, drawing, editing) and only write own UI, or you may try to embed root React component of our interface into your application.