Installation Instructions

You can bypass the installation steps below by using Windows or Mac OS Installer of OpenCRAVAT.

Windows users: Download the latest .exe installer from GitHub

Windows Defender may indicate that it prevented an unknown application from running. If so, click the ‘More’ link on the message and then select ‘Run Anyway’. An “OpenCRAVAT” icon will be created on the Desktop and the Start Menu. Double-click the OpenCRAVAT icon to launch the web interface.

Mac users: Download the latest .pkg installer from GitHub

Right-click the downloaded file and select “Open”. Your Mac may ask you to confirm stating that the file was not downloaded from the App Store. Click “Open” button. Just double-clicking the installer may not enable this “Open” option.

Installing with pip

To install Open-CRAVAT you need Python 3.6 or newer. If you do not have Python 3, install it using an installation file provided by

Installing with Conda

Open-Cravat is provided as a Conda package through BioConda. Ensure that Conda is configured to use the BioConda channel, and install Open-Cravat as following:

# One-Time configuration to add the bioconda channel
conda config --add channels defaults
conda config --add channels bioconda
conda config --add channels conda-forge
conda config --set channel_priority strict

# Install Open-Cravat
conda install open-cravat
oc version


On linux, OpenCRAVAT is installed with the ‘open-cravat’ pip package. The steps for installation are different if you are installing for a single user, or all users system.

For a single user, we recommend the following methods.

  1. Use a virtual environment such as venv, or conda.

  2. Install to your home directory with pip3 install --user open-cravat

To install for all users, run pip as root with sudo pip3 install open-cravat. Pip will place the oc executable at a location accessible to all users, probably /usr/local/bin. All users will be able to use oc to run jobs. However, only root will be able to install modules, or launch the gui. Unfortunately, in many systems /usr/local/bin is not in root’s path. This can be solved by linking oc to a directory in root’s path, such as sudo ln -s /usr/local/bin/oc /usr/bin/oc.


We recommend installing Python 3 using the installation file provided at instead of using any other manual way of installing Python 3. After installing Python 3, a new terminal should be opened and used in executing the below commands.


To install the openCRAVAT python package, run the following command in a command line shell.

pip3 install open-cravat

If you use Windows and if your Python 3 is installed inside of a system-level folder such as “C::raw-latex:`Program `Files”, you may have a problem with running open-cravat without the admin privilege. In this case, we recommend installing Python 3 outside of system-level folders and then installing open-cravat.

PyYAML error: If you experience an error such as ERROR: Cannot uninstall 'PyYAML'. It is a distutils installed project and thus we cannot accurately determine which files belong to it which would lead to only a partial uninstall., use a virtual environment or try pip install --ignore-installed PyYAML.

For Ubuntu: pip3 provided by apt does not install executables properly. We recommend the following steps before proceeding. sudo apt remove python-pip if pip3 has already been installed with apt. Then wget and sudo python3 For system-wide use of OpenCRAVAT, the following suggestion works: Install OpenCRAVAT with sudo (sudo pip3 install open-cravat), Use sudo for oc module install/uninstall/update and oc gui. Use without sudo for oc run and oc gui job.sqlite commands.

Install Base Components

OpenCRAVAT needs a set of base modules to function (converters, a mapper, an aggregator and reporters). The mapper has a full GRCh38 gene set database so will require 2GB of space to install.

oc module install-base

Running the web interface

Once the base install is complete, the OpenCRAVAT browser based GUI interface can be used to install annotators, submit, and view jobs. This can be done by following the instructions in the GUI usage page.

Alternatively, users who prefer a command line interface should proceed by reading the following sections.

Install Annotators

OpenCRAVAT is a modular system that enables users to install and run just those variant analysis modules that are relevant to their study. The OpenCRAVAT Store includes a wide variety of variant interpretation modules developed by both CRAVAT team members and by other members of the variant interpretation community. For example the grasp annotator provides GWAS associations for variants and vest provides scores that predicts pathogenicity. Annotators for both coding and non-coding variants are available.

To search the CRAVAT Store to see available annotators run:

oc module ls -a -t annotator

The result of this command is a list of available annotators that also indicates which annotators you have installed.



Latest version


Installed version




















To get a more detailed description for a module of interest, type:

oc module info <annotator name>

When you identify an annotator in the list above that you wish to use, install it with the command:

oc module install <annotator name>

For example: oc module install clinvar

Only those annotators that are useful for your study need to be installed. Many annotators include large reference data sets, which will take time to install and may require substantial disk space. A progress bar will indicate the portion of the data which has been downloaded as it runs.

Updating Annotators

Update all modules with oc module update. To update specific modules use oc module update [module1] [module2] As mentioned above, the oc module ls -a command will indicate if the version of a module installed locally is out of date.

Moving Modules Directory

The OpenCRAVAT pip package will normally be installed in your local Python site-packages directory. All of the additional annotators and other CRAVAT modules will be installed in the ‘modules’ directory of the OpenCRAVAT package. By default, this will be in the CRAVAT package in the site-packages directory. Occasionally the size of the module reference data is an issue on the drive where the Python site-packages are stored. If this is the case, you can relocate the CRAVAT modules directory to a different drive.

Use oc config md to see where modules are currently stored. To change the modules directory, copy data from the old modules directory to the new one, then use oc config md [new directory] to point OpencRAVAT to the new directory.

Installing Viewer Widgets

When OpenCRAVAT analysis is complete, one method of exploring the results is to use the graphical interactive results viewer oc gui. The oc gui program needs visualization widgets related to the annotators used in the CRAVAT analysis. In the future the viewer will be augmented to dynamically obtain all required viewer widgets. In the interim, if you plan to use the oc gui program to explore results, you need to install the appropriate viewer widgets. The base widgets for each annotator have a name that matches the annotator name with a ‘wg’ prefix. For example the cosmic annotator results are displayed with the wgcosmic viewer widget. To see available widgets:

oc module ls -a -t webviewerwidget

To install a widget:

oc module install widgetName

Multiuser GUI support

OpenCRAVAT’s web interface can have multiple user accounts by installing open-cravat-multiuser package (from OpenCRAVAT 1.6.0). See this page for the details of using multiple user accounts with OpenCRAVAT.