JupyterLab 4 Data Science Workstation
JupyterLab 4.6.4 on the Jupyter Docker Stacks scipy-notebook image, with pandas, NumPy, SciPy, scikit-learn, matplotlib, seaborn, statsmodels and Dask preinstalled. This template deploys one password-protected service with your whole home directory on a persistent volume.
What you get
- JupyterLab 4.6.4 on Python 3.13 (scipy-notebook image, weekly build 2026-10-05)
- A Railway volume at /home/jovyan holding notebooks, settings, pip installs and conda environments
- Login with a generated JUPYTER_PASSWORD and a JUPYTER_TOKEN for IDEs and scripts
- pip install goes to ~/.local on the volume, so packages survive upgrades
- Jupyter restarts itself inside the container if you shut it down from the menu
Good for
- Exploratory data analysis and visualisation
- Training and evaluating scikit-learn and statsmodels models on data too big for a laptop session
- Querying databases on your Railway private network from notebooks
- A remote Jupyter kernel for VS Code or PyCharm
Cost
Measured: about 110 MB for the Jupyter server and about 160 MB with one idle kernel. Cost on Railway then follows the RAM your kernels hold, since each kernel keeps its data in memory.
Services in this template
- JupyterLab · from
baranberkay96/jupyterlab-railway
Questions people ask
How do I log in?
Open the service URL and enter JUPYTER_PASSWORD from the Variables tab. To connect an IDE, use the server URL with ?token= followed by JUPYTER_TOKEN, or send the header Authorization: token {JUPYTER_TOKEN} from scripts.
Do my packages and notebooks persist?
Yes, everything under /home/jovyan is on the volume, including pip installs in ~/.local and conda environments. System packages installed with sudo apt-get (needs GRANT_SUDO=yes) reset on redeploy.
How do I keep memory use down?
Each open kernel keeps its data in RAM. Set JUPYTER_KERNEL_CULL_IDLE_TIMEOUT in seconds to shut down idle kernels automatically.
Can a team share one instance?
It is a single-user server, so keep one replica. For a team, deploy one instance per person.
How do I upgrade?
Change the ARG BASE_IMAGE tag in services/jupyterlab/Dockerfile to a newer weekly tag and redeploy. Packages in ~/.local are tied to the Python minor version, so reinstall them after a Python upgrade.
Full template README
Deploy and Host JupyterLab with Railway
JupyterLab 4 in your browser with a ready Python data stack: pandas, NumPy, SciPy, scikit-learn, matplotlib, seaborn, statsmodels, Dask and more. Your notebooks, installed packages, conda environments and settings live on a persistent volume, and the server is protected by a generated password and API token.
About Hosting JupyterLab
A notebook server needs more than a running container. It needs storage that survives redeploys, authentication that is on from the first request, and a way to add packages without rebuilding the image. This community template runs the official Jupyter Docker Stacks scipy-notebook image at a pinned weekly build, mounts the whole /home/jovyan directory on a Railway volume, and hashes the login password at start so it never appears inside notebooks or terminals. pip install and new conda environments go to the volume, so they are still there after an upgrade. Jupyter restarts itself inside the container if you shut it down from the menu.
Common Use Cases
- Exploratory data analysis and visualisation with pandas, matplotlib, seaborn and Altair
- Training and evaluating scikit-learn and statsmodels models on data too big for a laptop session
- Querying databases on your Railway private network from notebooks (SQLAlchemy included)
- A shared, always-on scratchpad for scripts, cron-like experiments and teaching
- A remote Jupyter kernel for VS Code or PyCharm
Dependencies for JupyterLab Hosting
- A Railway volume mounted at
/home/jovyan(included) - No external services or API keys are required
Deployment Dependencies
- Jupyter Docker Stacks: https://jupyter-docker-stacks.readthedocs.io/
- scipy-notebook image tags: https://quay.io/repository/jupyter/scipy-notebook?tab=tags
- JupyterLab documentation: https://jupyterlab.readthedocs.io/
- Jupyter Server security (passwords and tokens): https://jupyter-server.readthedocs.io/en/latest/operators/security.html
Implementation Details
| Service | Source | Purpose |
|---|---|---|
| JupyterLab | Dockerfile based on quay.io/jupyter/scipy-notebook:2026-10-05 (JupyterLab 4.6.4, Python 3.13) |
Web UI and kernels on port 8888, healthcheck /api, volume at /home/jovyan |
First login
- Open the service URL and enter the value of
JUPYTER_PASSWORDfrom the service's Variables tab. - Keep your work under
/home/jovyan(for example~/work); everything there persists. - To connect an IDE, use
https://{your-domain}/?token={JUPYTER_TOKEN}as the existing Jupyter server URL. Scripts can send the headerAuthorization: token {JUPYTER_TOKEN}.
Installing packages: %pip install {package} in a notebook or pip install {package} in a terminal installs into ~/.local on the volume (PIP_USER=1). For separate environments run mamba create -n {name} python=3.12 ipykernel, then ~/.conda/envs/{name}/bin/python -m ipykernel install --user --name {name}; both the environment and the kernel persist. GRANT_SUDO=yes enables sudo apt-get, but system packages reset on redeploy.
Saving memory: each open kernel keeps its data in RAM. Set JUPYTER_KERNEL_CULL_IDLE_TIMEOUT (seconds) to shut down idle kernels automatically.
Scaling: this is a single-user server, so keep one replica. Railway grows RAM and CPU vertically up to your plan limit. For a team, deploy one instance per person.
Pinning and upgrades: the base image is set by ARG BASE_IMAGE in services/jupyterlab/Dockerfile. Docker Stacks publishes a new date tag every week; change the tag and redeploy. Packages in ~/.local are tied to the Python minor version, so after a Python upgrade reinstall them.
Why Deploy JupyterLab on Railway?
Railway gives the notebook server an HTTPS URL, a persistent volume and private networking to your databases without any server administration. You pay for the RAM and CPU your kernels actually use, and you can resize or upgrade the image in minutes.
Guides for this template
View JupyterLab 4 Data Science Workstation on railway.com
Similar templates
| Template | Category | DeploymentsDeploys |
|---|---|---|
| 9Router Self-hosted AI router for Claude, Gemini, Cursor & OpenAI tools. | AI/ML | 9,734 |
| flowise-railway Flowise - low-code LLM apps builder | AI/ML | 8,512 |
| LibreChat Open-source ChatGPT clone: multi-model, agents, search, multi-user auth | AI/ML | 6,762 |
| Hermes Agent by Nous Research Template for Hermes Agent by Nous Research | AI/ML | 6,413 |
| OpenClaw [Homebrew support] The AI that actually does things | AI/ML | 5,167 |
| Hermes Agent Self Host Hermes Agent. Autonomous AI agent w/ self-improving, tools & more | AI/ML | 4,092 |