Skip to content

Getting started

The following sections will get you started with deploying your own Neurobagel node, a graphical query tool, and a local federation API (everything in blue in the picture below) that lets you search across the data in your node and in public Neurobagel nodes.

Neurobagel node

To prepare your Neurobagel node for production use (i.e., for local or other users), and to configure your deployment according to your specific needs, refer to the detailed Configuration documentation.


Neurobagel tools are provided as Docker containers and are launched with Docker Compose.

Don't install Neurobagel tools directly on your machine

Please only use the docker images provided by Neurobagel (or the third party providers Neurobagel relies on) and only launch them with our provided docker-compose.yml recipe.

Do not install GraphDB locally on your computer, as doing so can interfere with the deployment of the Neurobagel tools.

docker and docker compose

If you haven't yet, please install both docker and docker compose for your operating system:

  1. Install Docker Desktop on Windows. This will install both docker and docker compose.

  2. We strongly recommend also installing Windows Subsystem for Linux to get a Windows-supported Linux installation for a more seamless Neurobagel deployment experience. Simply follow these instructions to make your existing Docker Desktop installation (including Docker and Docker Compose) available when running WSL.

Install Docker Desktop on Mac. This will install both docker and docker compose automatically.

Linux is the only supported OS

We test and deploy on Linux and ensure that our deployment instructions work on Linux systems.

We also try to provide docs and help for different architectures, but as a small team with limited resources we won't be able to help you debug Operating System specific problems.

Because we rely on some modern features of these tools, please make sure you have at least the following versions on your machine:

  • docker engine: v20.10.24 or greater
    docker --version
  • docker compose: v2.0.0 or above
    docker compose version

The Neurobagel node deployment recipe

The neurobagel/recipes repository on GitHub contains our official Docker Compose recipe and template configuration files for setting up a local Neurobagel node.

  1. Clone the GitHub repository to your machine and navigate to it
    git clone
    cd recipes
  2. Make copies of the template configuration files to edit for your deployment (do not edit the templates themselves)
    cp template.env .env
    cp local_nb_nodes.template.json local_nb_nodes.json
  3. Change NB_API_QUERY_URL in the .env file

    You must replace the placeholder value for NB_API_QUERY_URLin the .env file

    with the address of the machine you are going to deploy Neurobagel on.

    • If you are deploying Neurobagel for yourself or just to try things out locally, you can use: NB_API_QUERY_URL=http://localhost:8080 where :8080 is the default port for the federation API.
    • If you are deploying Neurobagel on a server for other users, you must use the IP (and Port) or URL that your users will access the server with.


This is the minimal configuration you need to make before you can launch Neurobagel. In most cases, and especially when you are deploying Neurobagel for other users, you will have to make additional configurations.

Please refer to our detailed documentation for a complete overview of configuration options.

Launch Neurobagel

Once you have edited the .env and optionally the local_nb_nodes.json file, you can launch your own Neurobagel node with the following docker compose command:

docker compose --profile full_stack up -d

this will:

  • pull the required Docker images (if you haven't pulled them before)
  • launch the containers for the Neurobagel services using the full_stack service profile
  • automatically set up and configure the services based on your configuration files
  • automatically upload data to the Neurobagel graph (by default, it will upload an example dataset we have provided for testing)

You can check that your docker containers have launched correctly by running:

docker ps
and you will want to see something like this to show all 4 services running:
 docker ps
CONTAINER ID   IMAGE                              COMMAND                  CREATED         STATUS         PORTS                                                 NAMES
d5e43f9ff0c2   neurobagel/federation_api:latest   "/bin/sh -c 'uvicorn…"   8 seconds ago   Up 8 seconds>8000/tcp, :::8080->8000/tcp             recipes-federation-1
f0a26d0ea574   neurobagel/api:latest              "/usr/src/api_entryp…"   8 seconds ago   Up 8 seconds>8000/tcp, :::8000->8000/tcp             recipes-api-1
d44d0b7359c8   ontotext/graphdb:10.3.1            "/usr/src/neurobagel…"   8 seconds ago   Up 8 seconds>7200/tcp, :::7200->7200/tcp, 7300/tcp   recipes-graph-1
29a61a2d83de   neurobagel/query_tool:latest       "/bin/sh -c 'npm run…"   8 seconds ago   Up 8 seconds>5173/tcp, :::3000->5173/tcp             recipes-query_federation-1

The docker-compose.yml recipe provides additional service profiles for different deployment use cases (e.g., if you do not need to set up local query federation). Please refer to our service profile documentation for details.

Next steps

🎉 You are now the proud owner of a running Neurobagel node. Here are some things you can do now: