Links

Installation & upgrade

How to install Giskard on your machine, set up your Python backend and upgrade Giskard
Cloud installation:

Requirements

To install Giskard, you need a Linux or macOS machine with:
  • By default Giskard uses 2 TCP ports: 19000 and 40051, make sure they're open on the machine where Giskard is installed
  • git (download)
  • docker (download).For an easy installation of Docker you can execute:
    curl -fsSL https://get.docker.com -o get-docker.sh
    sudo sh get-docker.sh
Run the following commands to install Giskard on your server
git clone https://github.com/Giskard-AI/giskard.git
cd giskard
docker compose pull && docker compose up -d --force-recreate --no-build
  • If you have an error because of the rate limit (toomanyrequests: Rate exceeded), please re-execute the docker-compose command line once again.
  • If you have an older version of docker, you'll need to usedocker-compose instead of docker compose
Once the docker-compose starts all the modules, you'll be able to open Giskard at http://localhost:19000/
Since the backend container may take some minutes to load, please wait a bit and refresh the webpage http://localhost:19000/
To log in, the default credentials are Login: admin / Password: admin
You're all set to try Giskard in action. Upload your first model by following the Upload your ML model & data tutorial.

Upgrade Giskard to the latest version

In order to upgrade Giskard to the latest version, please run the following in your Giskard distribution directory
git pull
docker compose down && docker compose pull && docker compose up -d --force-recreate --no-build
  • The browser may keep using an old version of Giskard UI due to caching. If there are issues after running an upgrade try to hard refresh Giskard page by pressing Ctrl + Shift + R or Command + Shift + R
  • If you installed in Giskard additional Python libraries or a new Python version, you will need to reinstall them. Please refer to configuration.

Troubleshooting

If you encounter any issues, join our Discord on our #support channel. Our community will help!