Jupyter Lab

JupyterLab is a web-based interactive development environment for notebooks, code, and data, backed by remote compute. This image bundles JupyterLab on top of Miniconda, served straight to the browser — no desktop or VNC.

Configuration

name: jupyter
tags:
    - analytics
steps:
    - name: app
      platform: docker
      mode: parallel
      image: ghcr.io/dxflow-ai/jupyter:latest
      volumes:
          - name: volume
            host: ./volume
            container: /volume
      ports:
          - name: web
            host: "8888"
            container: "8888"
      env:
          - WORKING_DIR=
      resources:
          cpu: "4"
          memory: 8G

[volume]
app.volume = ./volume

[port]
app.web = 8888

[env]
app.WORKING_DIR =

[resource]
app.cpu = 4
app.memory = 8G

{
    "arch": ["amd64", "arm64"],
    "image": "ghcr.io/dxflow-ai/jupyter:latest",
    "version": "4.2",
    "minimum": {
        "cpu": 4,
        "memory": "8G",
        "storage": "50G"
    }
}

Usage

1. Deploy

dxflow workflow create --identity jupyter jupyter.yml

# Start with defaults, or open a specific working directory
dxflow workflow start jupyter
dxflow workflow start jupyter \
    --override env.app.WORKING_DIR=projects/analysis

2. Open the notebook

Open your browser at http://localhost:8888. JupyterLab opens on the working directory and needs no token — keep the port private and reach it through the platform's authenticated proxy.

3. Persist data

Notebooks and data live under /volume, so your work survives restarts — mount a local directory there to keep it.

Notes

• WORKING_DIR is resolved under /volume (empty opens /volume). Set it to a subpath like projects/analysis to open straight into a project.
• The token is disabled (--NotebookApp.token=''); JupyterLab is meant to sit behind the platform's authenticated proxy, so do not expose port 8888 directly to the internet.
• Miniconda is at /opt/miniconda and on the PATH — use conda and pip from a notebook terminal to add libraries such as numpy, pandas, scikit-learn, or extensions like jupyterlab-git.