EchoVision Wish List: Scene Description and Questions

Last updated:

Describing what is in front of you, and answering your questions about it.

Part of the EchoVision Wish List. See also how we decide on ideas.

6 ideas: 1 in testing, 1 in progress, 1 under review, 2 released and 1 merged.

In testing

SCENE-05: Scan barcodes and QR codes

  • Status: In testing. An early barcode scanner is in beta testing.
  • What owners asked: Could the glasses scan barcodes and QR codes for product information and expiry dates?

In progress

SCENE-03: Add people to face recognition from my photos

  • Status: In progress. Face recognition is in beta since June 2026, in scene description only. Accuracy needs more work before a wider release.
  • What owners asked: Is it envisioned that a facial recognition "memory" of images of people could be integrated i.e. This is a picture of aunt Sally and cousin Betty.
  • Related: GEN-05: Recognize people I know

Under review

SCENE-06: Read clocks, dials and screens more reliably

  • Status: Under review. These work today, but haven't been specifically tuned yet.
  • What owners asked: Could it read analog clocks, dials, appliance displays and computer screens more reliably?
  • Related: LIVE-11: Read signs, labels and displays aloud

Released

SCENE-02: Describe what I'm holding first

  • Status: Released. Model tuned to give priority to products held in hand.
  • What owners asked: If we're holding something like a food product right in the foreground, should the glasses assume we want priority given to describing it without having to ask?
  • Related: LIVE-03: Describe what I'm holding first, in Live AI

SCENE-04: One press to describe, press and hold to ask

  • Status: Released. New AI button behavior in October 2026, with two settings: "Long Press Scene Description" (turn it off for the old behavior) and "Follow-up Questions".
  • What owners asked: Could one press give an instant description, and press-and-hold let me ask a question? And could I keep the old behavior and ask follow-up questions?

Merged

SCENE-01: Warn when text is blank, upside down or sideways