Overview

JupyterLab is a web-based environment for working with notebooks, code, and data. It is best suited to people who want an adaptable workspace for exploratory computing and are comfortable installing software locally. Its breadth is a strength, but extension permissions and the lack of a hosted service make it less convenient for teams seeking a managed, centrally assessed platform.

Key features

A workspace beyond notebook cells

Tabs and splitters let users keep notebooks, text editors, terminals, and custom components together. JupyterLab also opens and edits Markdown, images, CSV, JSON, HTML, LaTeX, and PDF, and supports the .ipynb format used by classic Jupyter Notebook. That range suits work that moves between code, data, and written analysis; exports include HTML, Markdown, LaTeX/PDF, and reStructuredText.

Interactive computing

Code consoles offer scratch space for experiments, and kernel-backed documents can run code in formats such as Markdown, Python, R, and LaTeX. Supported languages include Python, R, C++, Julia, GNU Octave, Ruby, and Scheme. This flexibility serves varied technical workflows, though visualizations using matplotlib or Plotly may need additional packages for interactive plots or widgets.

Extensions, collaboration, and trust

Extensions can add themes, file viewers and editors, output renderers, menus, shortcuts, and settings. Real-time collaboration is available by installing the jupyter_collaboration extension. The tradeoff is meaningful: extensions run JavaScript in the browser and can execute arbitrary code on the server, kernel, and client browser, so users should install only trusted ones. Notebook JavaScript and HTML outputs from other machines are sanitized and remain hidden until the notebook is trusted, a useful safeguard when opening unfamiliar files.

Pricing

JupyterLab is free, open-source software under a modified BSD license. The JupyterLab plan costs 0.00 USD per free and requires local installation with pip; the installation instructions also provide a Homebrew command for macOS and Linux. There is no paid tier to trade up to, but users must provide and manage their own environment rather than rely on JupyterLab as a hosted service.

That deployment model also matters to organizations: Project Jupyter says it does not provide the software as a service and cannot complete vendor assessment questionnaires such as SOC 2 or SOC 3. Teams that require a vendor-operated service or those assessments should look elsewhere.

Platforms

JupyterLab supports Linux, macOS, self-hosted deployments, and web access. Its execution model is local, so the browser-based interface does not remove the need to install and run the software in an environment the user or organization manages.

Who it's for

JupyterLab fits data science, scientific computing, computational journalism, and machine-learning workflows where interactive, exploratory work matters more than a turnkey hosted setup. It is a strong choice for users who want notebooks and other working documents side by side, can manage a local installation, and are prepared to vet extensions. It is less suitable for buyers who need a vendor-provided service or vendor security assessments.

Pros and cons

Pros: A single, rearrangeable workspace brings notebooks, terminals, editors, and custom components together, helping keep exploratory work in context.
Pros: Broad format and language support, plus export options, makes it useful for workflows that span code, data, and documentation.
Pros: The software is free and open source, with real-time collaboration available through an extension.
Cons: Local installation and execution place setup and environment management on the user or organization.
Cons: Extensions can run arbitrary code across the server, kernel, and browser, making extension trust a security decision rather than a cosmetic one.
Cons: Project Jupyter does not provide JupyterLab as a service or complete SOC 2 or SOC 3 vendor questionnaires, which can rule it out for some organizations.

Alternatives

Google Colab is worth considering if a browser-accessible notebook service is preferable to JupyterLab's local execution; its free plan has no guaranteed access to compute units and limited GPU access.

Deepnote may suit readers comparing a free notebook option with editor and project caps: its free plan allows up to 3 editors and 5 projects.

Datalore is another hosted notebook option to compare, with a free plan that caps CPU S machine time at 120 hours and cloud storage at 10 GB.

Kaggle Notebooks is a free alternative for readers who can work within session limits: CPU/GPU sessions last up to 12 hours and TPU sessions up to 9 hours, while free P100 GPU and TPU v3-8 availability can be limited or queued.

Hex offers a free Community plan with small compute, making it an alternative for readers willing to work within that plan's compute scope.

Observable is another option for readers comparing free plans; its Free plan includes few tokens per day and 1 GB per month of file attachments, with embeds carrying a watermark.

marimo is another free alternative.

Saturn Cloud is a paid alternative for readers considering hosted computing; its Hosted plan is pay-as-you-go and billed hourly.

Browse Notebook Software or Data Science Platforms for more options.

Verdict

Choose JupyterLab if you want a free, flexible environment for interactive notebooks and exploratory computing, and are willing to install and manage it locally. Its workspace and format range are compelling for technical users; look elsewhere if you need a vendor-hosted service, formal vendor security assessments, or a setup that avoids extension-risk decisions.

	JupyterLab	RunPy
Free plan	Yes	Yes
Free trial	No	No
Paid from	Free	£8/mo
Open source	No	No
Platforms	Linux, macOS, self-hosted, Web	—
Free plan	Yes	Yes
Private projects	—	Yes
Full-stack runtime	—	No
Deployment options	—	none
Runtime languages	—	2 languages
Sharing controls	—	link_access

JupyterLab vs RunPy

JupyterLab

RunPy