Your robot can read now — on-device text recognition
Point the camera at a sign, a label, or a door plate and the robot reads it. Recognition runs on the robot itself, so no image ever leaves the machine.
The Robot Reads What It Sees
Cameras let a robot see shapes, but most of the useful information in a building is written down — room numbers, equipment labels, shelf tags, warning signs, meter readings. Until now the robot could look straight at a door plate and still not know which room it was standing in.
Text recognition is now part of the perception stack.
What's New
- Reads signs, labels, and plates — Any text in the camera frame comes back as text the robot can reason about
- Chinese and English together — Mixed-language signage is handled in a single pass, no need to pick a language first
- Runs on the robot — Recognition happens on-device; frames don't need to be uploaded anywhere to be read
- Feeds straight into the AI — The recognised text becomes something the agent can act on: "go to the room labelled 302", "tell me what this warning says"
What You Can Do With It
Ask the robot to find a specific room by its door plate, read a meter and report the number, check that a shelf label matches what's on the shelf, or read out a notice for someone who can't reach it. Anywhere a person would have to walk over and squint at a label, the robot can now do it instead.