Help Centre / Features / AI video annotation (Beta)

AI video annotation (Beta)

Updated just now · 8 views ·

Video annotation lets you drop in a phone or GoPro recording of a process and get a first-draft map back. This is a beta feature - expect the accuracy and speed to keep improving.

How it works

  1. Upload a process video (max 15 minutes, mp4/mov).
  2. MapVS runs pose estimation and object detection to segment the video into steps.
  3. For each step you get a suggested name, cycle time, and VA/NVA classification.
  4. Review, edit, and merge into a map.

What works well today

  • Repetitive manual assembly at a fixed station.
  • Well-lit videos with the operator clearly in frame.
  • Segments of 30 seconds to 5 minutes.

What doesn't yet

  • Multi-operator scenes (accuracy drops significantly).
  • Long walking segments with no clear station transitions.
  • Video with heavy occlusion or dark environments.

Quota

Starter: 5/month. Pro: 20/month. Team: 50/month. Enterprise: unlimited. All Beta - subject to change as the feature matures.

Was this article helpful?

0 found this helpful