Core update

Companion Vision in Omni AI Companion

Companion Vision lets you share a visual window into the moment you are in. It is useful when pointing, showing, or looking together would communicate more naturally than describing every detail in text.

What changed

Context you can point toward

The visual experience gives your conversation something concrete to notice: an object on a desk, a view outside, a handwritten page, or a small detail you want to discuss.

Seeing is still a conversation

The feature is designed for shared attention, not silent certainty. Ask questions, correct what was missed, and explain why the detail matters instead of assuming an image tells the whole story.

Why this feature matters

When Rosa is rearranging a room, she can show two possible corners and talk through the feeling she wants—calmer, brighter, or less crowded—before choosing one.

A student can point to a diagram that feels confusing and ask for a step-by-step explanation, then follow up with the exact part that still does not make sense.

Feature examples and use cases

  • Choose what to share: Frame the object, view, or page that is relevant and remove unrelated details from the moment if you can.
  • Ask a visual question: Say what you want noticed—comparison, description, organization, or a closer look—instead of only showing the image.
  • Stay in dialogue: Confirm observations and add context when the visual scene could be ambiguous.

How to use it

  1. Open Companion Vision: Launch Companion Vision in Omni and follow the on-screen request to begin.
  2. Allow camera access: Grant camera permission when requested, then position the view around the detail you want to share.
  3. Explore together: Ask about the scene, clarify what matters, and stop sharing when the visual context is no longer needed.

Tips for a better first try

  • Use steady lighting and a close, simple frame.
  • Name the important detail if the scene contains several possible subjects.
  • Avoid showing private documents, other people, or sensitive surroundings unless you have a clear reason and permission.

Limitations and things to know

  • Visual interpretation can miss small text, context, colors, or details; treat it as a conversation aid rather than proof.
  • Camera access and the experience may depend on your device permissions and current app support.

Availability

Open Companion Vision to see whether it is available on your account and device. Permission prompts and current access are shown in the experience; do not assume support on every platform.

Frequently asked questions

What should I show first?

Start with one relevant object, page, or view and explain what you want to understand about it.

Can it identify everything in a scene?

No. Visual context can be incomplete or ambiguous. Ask focused questions and verify important observations yourself.

How do I stop sharing the view?

Use the on-screen control to end or leave the visual experience, and review your device permission settings if needed.

Open Companion Vision in Omni

Related Omni updates

Browse every Omni feature update · Explore step-by-step tutorials

Omni AI
Omni AI
Your AI Companion