# October 12, 2022

> The internal architecture of the neural network used for our Computer Vision server has been improved, and now yields better results when used in the context of automating text input boxes.

## Model version: 2022.10.1.3

## Improvements

The internal architecture of the neural network used for our Computer Vision server has been improved, and now yields better results when used in the context of automating text input boxes.

To speed up the migration to the new Computer Vision endpoint and make it effortless for our customers, we have implemented an automatic nearest geo location endpoint selector for `https://cv.uipath.com` based on the request. This change automatically improves latency.

![docs image](https://dev-assets.cms.uipath.com/assets/images/ai-computer-vision/ai-computer-vision-docs-image-180364-4830b22e-444d6c59.webp)

## Bug fixes

Fixed an issue which caused auto-anchors to be missing at design time or when using the **Capture All Elements** feature.
