Get VisionEval Here

Version 3.1. 468 Mb download! Packaged for R 4.3.2

The link above will download a .zip file containing the following:

Other R versions for the latest release are available here (Scroll down to the Assets section). Installers for older releases can also be found on Github.

Developers can find the latest development branch in the development branch of the VisionEval-Dev repository.

Install for Windows

You can find complete instructions in the VisionEval Getting Started documentation You will first need to install a supported version of R and (if you can RStudio).

After installing R (see Requirements below) and downloading the corresponding VE Installer for that R version from the link above, unzip the folder to the destination folder of your choice.

To complete the installation and start VisionEval with RStudio, simply:

You can also start VisionEval in the standard R GUI by double-clicking launch.bat. Please see further setup instructions if launch.bat seems not to work.

Getting Started

Once you have been welcomed to VisionEval, you can follow the instructions under “Editing and Running Models” on the Getting Started page. Your destination folder contains everything you need from the VisionEval “sources” folder.


If the above installation steps did not succeed, ensure that you have downloaded the appropriate version of VisionEval to match the version of R that you have installed.


The current version of VisionEval is built for recent versions of R. If you have a different version installed and cannot install one of the supported versions, please contact Jeremy Raw. You can get R here.


We recommend using RStudio to work with VisionEval. RStudio is particularly recommended if you plan to clone and explore the Visioneval source code from GitHub, and it is very popular among regular R users.

Mac/Linux users

VisionEval can be installed from source as well. Contact Jeremy Raw for more information on source installations.


Questions about VisionEval installation can be directed to Jeremy.Raw or Scott.Smith at, or more generally to

The installers were built using the process described here on GitHub