Notebook Services
Browser-based notebook platforms can be useful for teaching, exercises, data analysis, and machine-learning experiments without requiring students to configure a local environment.
These are external services and are not operated by AISL or FHNW. Free plans, compute quotas, hardware availability, privacy practices, and terms can change. Check the provider's current conditions before relying on a service for a course or project.
Noto (EPFL)
Noto is EPFL's centralized JupyterLab platform for education. It runs entirely in the browser and supports Python, R, C, JavaScript, and Octave.
Free resources: Standard Noto provides free computing for teaching and learning. EPFL currently documents a maximum allocation of 4 GB RAM and 2 CPUs per user. GPU-enabled GNoto resources are separate and provided on request for suitable courses.
Access: FHNW users can sign in with a SWITCH edu-ID linked to their FHNW organizational identity. Noto is open to users from Swiss universities in the SWITCH network.
Best for: Course exercises, demonstrations, assignments, and lightweight Jupyter-based teaching with almost no setup.
RenkuLab
RenkuLab is a collaborative data-science platform developed by the Swiss Data Science Center. It connects code, data, compute, and browser-based development environments such as Jupyter and VS Code.
Free resources: Creating projects and running sessions in the public compute tier is free. The available resource classes can change with the platform configuration. Public sessions currently auto-pause after about 2 hours of inactivity and are shut down after 2 weeks if not resumed.
Access: Accounts can be created using SWITCH edu-ID, GitHub, or ORCID.
Best for: Reproducible projects that combine notebooks, repositories, datasets, collaboration, and configurable compute.
Kaggle Notebooks
Kaggle provides hosted notebooks with free CPU and accelerator resources alongside datasets, models, competitions, and learning resources.
Free resources: Kaggle currently documents notebook sessions of up to 12 hours for CPU/GPU and 9 hours for TPU, with 20 GB of autosaved working storage. CPU sessions provide 4 CPU cores and 30 GB RAM. Free GPU access includes NVIDIA P100-class resources; the weekly GPU quota is typically 30 hours and can sometimes be higher depending on demand. TPU access is documented at up to 20 hours per week.
Access requirements: A Kaggle account is required. Phone verification is required for accelerator access; Kaggle may apply additional verification requirements to some newer features or accounts. Hardware and quotas are subject to availability.
MCP: Kaggle also provides an official remote Model Context Protocol (MCP) Server for accessing Kaggle resources from compatible AI clients and agents.
Best for: Data-science exercises, public datasets, machine-learning experiments, competitions, and free GPU/TPU notebook workloads.
Deepnote
Deepnote is a collaborative cloud notebook designed for data science, teaching, and team-based analysis.
Free Education plan: Students, teachers, and researchers can qualify using an institutional email address. Each member of an Education workspace currently receives unlimited Basic-machine hours with up to 5 GB RAM and 2 vCPU. Advanced machines and GPUs require a payment method, and the free Education plan does not include advanced compute, Scheduling, or Deepnote AI.
Access: Sign up and verify your FHNW school email address to request the Education plan.
Best for: Collaborative notebook teaching, group assignments, shared classroom workspaces, comments, and real-time editing.
Google Colab
Google Colab is Google's hosted Jupyter Notebook service with free CPU access and optional GPU and TPU runtimes.
Free resources: Colab provides free access to GPUs and TPUs, but does not publish fixed accelerator quotas or guarantee specific hardware. Limits, idle timeouts, GPU/TPU types, and availability change dynamically. Free notebooks can currently run for up to 12 hours, depending on availability and usage patterns.
Access: A Google account is required. Free accelerator access is subject to Colab's dynamic usage and anti-abuse limits.
Best for: Quick notebook experiments, teaching examples, Google Drive integration, and occasional GPU/TPU workloads where fixed capacity is not required.
Choosing a service
Noto
FHNW/Swiss university teaching
4 GB RAM, 2 CPUs per user; SWITCH edu-ID
RenkuLab
Reproducible research and project collaboration
Free public compute tier
Kaggle
ML exercises and accelerator workloads
CPU plus weekly GPU/TPU quotas
Deepnote
Collaborative classroom notebooks
Unlimited Basic-machine hours on Education plan
Google Colab
Quick experiments and examples
Dynamic free CPU/GPU/TPU access