[Feature Request - Nok] Interop between DLS+Geti+Scenescape

[jaisuthan]

✨ Feature description

Make DLStreamer, Intel Geti, and Intel Scenescape work together automatically, so a user doesn’t have to manually move videos, create projects, or handle models by hand.

💡 Solution description

• Video streams (RTSP / camera feeds) are:
  Captured once and reused for both training (Geti) and inference (DLStreamer)

• Geti projects (datasets, labels, training configs) are:
  Created automatically via API, possibly based on a template (e.g., “person detection”)

• After training:

  • Geti exports an optimized OpenVINO model
  • The model is automatically:
    Downloaded
    Stored
    Versioned

No manual model conversion or copying

---

🧠 Tip: You can upvote this request by reacting with 👍 to help prioritize it!

[saratpoluri]

@jangorecki1 : let us sync in the new year on how to distill this into the roadmap.

[saratpoluri]

Discussed with Product Management. Much of the API endpoints already exist in DLS-PS and Geti. SceneScape will do the engineering effort to stitch the workflow and highlight any gaps. No specific release commitment but target completion before EOY 2026.