diff --git a/computer-use-demo/README.md b/computer-use-demo/README.md index 2b896da8..ad96a3d4 100644 --- a/computer-use-demo/README.md +++ b/computer-use-demo/README.md @@ -157,6 +157,11 @@ docker run \ We do not recommend sending screenshots in resolutions above [XGA/WXGA](https://en.wikipedia.org/wiki/Display_resolution_standards#XGA) to avoid issues related to [image resizing](https://docs.anthropic.com/en/docs/build-with-claude/vision#evaluate-image-size). Relying on the image resizing behavior in the API will result in lower model accuracy and slower performance than implementing scaling in your tools directly. The `computer` tool implementation in this project demonstrates how to scale both images and coordinates from higher resolutions to the suggested resolutions. + +When implementing computer use yourself, we recommend using XGA resolution (1024x768): +- For higher resolutions: Scale the image down to XGA and let the model interact with this scaled version, then map the coordinates back to the original resolution proportionally. +- For lower resolutions or smaller devices (e.g. mobile devices): Add black padding around the display area until it reaches 1024x768. + ## Development ```bash