Paperspace Notebooks

Paperspace Notebooks is a hosted Jupyter workspace for data exploration and machine-learning development, pairing browser notebooks with CPU and GPU machines. It is a strong fit for developers and small teams that need shared projects and managed acceleration rather than local hardware. The free tier is useful for short experiments, while longer jobs and larger storage require a paid plan.

Overview

Notebooks runs Jupyter files in Docker containers, giving projects a hosted execution environment with fast container startup. You can launch a notebook from a browser, select a pre-built template or bring your own, and work in Python. That removes local environment and GPU setup from the workflow, although machine availability and capability depend on the plan you choose.

Shared persistent storage keeps files available across notebooks and supports ongoing team work. Teams can invite collaborators, assign permissions, fork public projects, and create notebooks. This is more suitable for coordinated development than an isolated notebook editor, but storage allowances are finite and paid plans charge for additional capacity.

Key features

Compute and machine-learning stack

GPU-enabled Jupyter machines are available from the browser, with higher tiers adding more capable options. The listed choices include free CPU (C4) and Free GPU (M4000) machines, plus P4000, P5000, RTX4000, RTX5000 and A4000 on Pro. Growth adds A5000, A6000 and A100-80G. That progression lets occasional users start cheaply and gives demanding model workloads a path to stronger hardware, but the most capable GPUs are reserved for the higher-priced tier.

Framework and tooling support covers TensorFlow, PyTorch, Keras, Scikit Learn, TensorRT, TensorBoard, ONNX, XGBoost, HuggingFace, RAPIDS and Fast.ai. The breadth suits common machine-learning pipelines; projects dependent on an unlisted framework should verify compatibility before committing.

Collaboration, data and automation

Real-time collaboration, permissions and project forking make shared notebook development practical. Pro and Growth subscribers can mount public or private S3-compatible sources, including DigitalOcean Spaces and AWS S3, directly into notebooks. The Paperspace API manages resources through HTTP requests, useful for scripted operations and repeatable administration.

Machine auto-shutdown can be configured from one hour to one week. Free-tier machines cannot exceed six hours, so unattended free workloads must finish within that window or be restarted. Security operations include round-the-clock threat monitoring and historical user and resource activity logs; datacenters are individually audited against SOC 1, SOC 2, PCI-DSS and ISO 27001 standards.

Pricing

Free

The Free plan costs nothing and includes 5 GB storage, Free CPU (C4) and Free GPU (M4000) machines. Free machines are restricted to Notebooks in private workspaces, and their maximum auto-shutdown timeout is six hours. It is appropriate for learning and bounded experiments, not persistent or long-running training.

Pro plans

Pro (user plan) costs $8 a month, billed at $0.0119/hr up to a maximum of $8.00/mo. Pro (team plan) costs $12 a month, billed at $0.0179/hr up to $12.00/mo. Both include 15 GB storage, charge $0.29/GB/mo for additional storage, and provide C4, M4000, P4000, P5000, RTX4000, RTX5000 and A4000 machines. The user plan suits an individual; the team plan is the clearer choice when shared ownership and collaboration are central.

Growth plans

Growth (user plan) and Growth (team plan) each cost $39 a month, billed at $0.0580/hr up to a maximum of $39.00/mo. Both include 50 GB storage, charge $0.29/GB/mo beyond that, and add A5000, A6000 and A100-80G machines to the Pro lineup. Choose Growth for heavier GPU workloads or larger projects; the identical price and machine list means the user-versus-team distinction should drive the choice.

Platforms

The hosted service is delivered through the web and offers an API, with support for Linux, macOS and Windows environments. The Paperspace App supports Windows 10+, OS X 10.13+ and Linux beta, adding hotkeys, drag-and-drop uploads and multi-monitor support. Standard support is free and ticket-based seven days a week; enterprise support by contract adds infrastructure assistance and customer success managers.

Who it's for

Choose Notebooks if you want browser-based Jupyter work with optional GPUs, shared storage and team controls. It is particularly well matched to solo developers moving beyond local hardware, small machine-learning teams, and organizations that need API-managed resources. Look elsewhere if you need unrestricted free runtimes, a self-hosted deployment, or a notebook tool without machine and storage charges.

Pros and cons

• Pros: Browser GPU notebooks avoid local accelerator setup.
• Pros: Docker environments, persistent shared storage and real-time collaboration support repeatable team projects.
• Pros: Pro and Growth S3-compatible mounts connect notebook work to public or private object storage.
• Pros: Growth includes A100-80G access for demanding training workloads.
• Cons: Free machines stop after six hours, limiting long unattended jobs.
• Cons: Storage is capped at 5 GB on Free and 15 GB on Pro, with $0.29/GB/mo overage on paid tiers.
• Cons: The strongest GPUs require the $39 a month Growth tier.

Alternatives

Browse Notebook Software for a wider directory of notebook tools.

Pick marimo for a free, open-source notebook available through web, desktop, API and self-hosted options.

Choose Apache Zeppelin when Apache-licensed, open-source software for Linux, macOS, web or self-hosting is the priority.

Deepnote is an alternative freemium workspace with a free plan, trial and limits of three editors and five projects.

Use Pluto.jl for a free, MIT-licensed open-source notebook.

SageMath fits users wanting free open-source software installable on unlimited computers.

Consider Datalore for a free cloud tier with 120 CPU S machine hours, 10 GB storage and two notebooks in parallel.

Dnote is a free, open-source option with an optional self-hosted server.

Choose JupyterHub when you want open-source Jupyter hosting on your own hardware or cloud.

Verdict

Paperspace Notebooks is a sensible choice for developers and small teams that need collaborative Jupyter projects with selectable GPU capacity and managed containers. Its main advantage is combining hosted acceleration, shared storage and team controls in one workflow. Look elsewhere if six-hour free-machine limits, paid storage overages or the need for self-hosting make those trade-offs unacceptable.