[FEATURE] Hope to add the feature of multimodal or vision understanding through IP camera

[guihaoqun]

Description

The ten framework can implement the local camera to obtain the video stream, but the reality is more about IP cameras，so Hope to add the feature of multimodal or vision understanding through IP camera

Severity

Critical

Additional Information

nothing

[plutoless]

could you pls describe the use case in more detail? how will you interact with agent? via the mic on ip camera?