Aiviro can be run locally at your computer or it can be deployed to a server with its orchestration system.

Local deployment

For running it locally, just run your main python script of the scenario you want to start.

Server deployment


Aiviro orchestration is deployed to a server with the help of docker & docker-compose. Therefore, you need to have docker & docker compose installed on your server.

  1. Connect to your server via ssh.

  2. Go to the directory where you want to deploy Aiviro. We recommend creating a new directory for it.

mkdir -p projects/aiviro  # create a new directory
cd projects/aiviro  # go to the directory
  1. Create and activate a python virtual environment.

python3 -m venv venv  # create a virtual environment
source venv/bin/activate  # activate the virtual environment
  1. Install Aiviro with Prefect extra dependencies.

pip install -i aiviro[prefect]
  1. Deploy Aiviro orchestration server. It contains server, database and web interface.

# deploying server with the default settings
aiviro-cli deploy server start --detach

# for information about custom deployment of the server, use --help option
aiviro-cli deploy server start --help
  1. Deploy Aiviro agent.

# deploying agent with the default settings
aiviro-cli deploy agent start --agent-name <your-agent-name> --git-url <repo-url-with-access-token> --label rdp --max-flows 1 --detach

# deploying agent with custom prefect project name
aiviro-cli deploy agent start --project-name <prefect-project-name> ...

# for further information about custom deployment of the agent, use --help option
aiviro-cli deploy agent start --help
  1. Add aiviro configuration file and set API credentials.

aiviro-cli config login --client-id <your-client-id> --api-key <your-api-key> --output prefect-data/<agent-name>/volume/
  1. Open the web interface in your browser, default url is http://localhost:8080. Select the Flow tab, Hello World flow and run it, in order to validate your installation.


Web interface with selected Flow tab


Detailed view of the Hello World flow

  1. To stop agent and server, use the following commands. You must be in the same directory where you deployed them.

# stop agent
aiviro-cli deploy agent stop --agent-name <your-agent-name>

# stop server
aiviro-cli deploy server stop

Successful deployment will look like this, you can check all running docker containers by running docker ps command.


Running docker containers

Register Flow

The deployed agent contains your git repository and therefore also your flows. However, you need to register them in the deployed environment, so that you can run them from the web interface. Follow these steps to register your flow:

# connect to the linux server
ssh <your-username>@<your-server-ip>
# enter the agent container
docker exec -it <agent-container-name> /bin/bash
# navigate to git repository directory
cd git-repo/
# register your flow
python flows/<your-flow-script>.py

Successful registration of the flow will look similar to this:


Successful flow registration

Update Agent

To update the agent, you have two options (soft and hard one). First step is always the same, you need to firstly stop the agent.


Keep in mind that by stopping or restarting any docker-container you’ll lose all temporary data stored in it. The initial state of the docker-container is defined by the image it was created from.

The soft way to update the agent is to start it using already existing .env file with the --update option:

aiviro-cli deploy agent start-env --agent-name <your-agent-name> --update

In case you will encounter any issues with the update, you can try to update the whole agent environment with agent start command:

aiviro-cli deploy agent start --reset-env --agent-name <your-agent-name> --git-url <repo-url-with-access-token> --label <label> --max-flows <number> --detach