Google Colab
Build
Google Colab A browser notebook for Python experiments. Start with a simple baseline, then add complexity only if it improves your test result.
Try this: Train a small classifier and publish the notebook with its dataset and random seed.
Explore Google Colab Hugging Face
Build
Hugging Face Explore model cards, datasets, and small interactive demos. Open weights do not automatically mean unrestricted use.
Try this: Compare the intended use and license of two models for the same task.
Explore Hugging Face Build
Google AI Studio A workspace for experimenting with Google’s generative models and multimodal inputs.
Try this: Evaluate the same prompt against a fixed set of cases and record failures.
Explore Google AI Studio Write
Claude A language assistant for drafting, coding, and analyzing documents. Check quotes and cited evidence against the original source.
Try this: Ask it to critique a methods section, then review every suggestion yourself.
Explore Claude Write
ChatGPT A conversational tool for exploring ideas, drafting explanations, and helping with code. Outputs need independent verification.
Try this: Use it to propose edge cases for a project, then run the tests.
Explore ChatGPT Zotero
Research
Zotero A reference manager that organizes papers, notes, and bibliographies. It helps make a publication’s evidence traceable.
Try this: Build a five-paper library and attach a one-paragraph methods note to each.
Explore Zotero PubMed
Research
PubMed Search biomedical literature, including publication metadata and links to available full text.
Try this: Compare a review paper with the original studies it cites.
Explore PubMed Research
AlphaFold Server A research service for predicting biomolecular structures with AlphaFold 3. Review the service’s current access terms.
Try this: Explore a public example and inspect confidence, not only the rendered structure.
Explore AlphaFold Server scikit-learn
Build
scikit-learn A Python toolkit for classical machine learning, preprocessing, and evaluation. Small, interpretable baselines are often valuable.
Try this: Compare logistic regression with a tree using the same train/test split.
Explore scikit-learn Build
MediaPipe Tools for on-device perception, including hand gestures and pose landmarks. Camera quality affects performance.
Try this: Try SurfAI’s gesture lab and document failures across lighting conditions.
Explore MediaPipe Write
Notion AI Workspace assistance for organizing, finding, and drafting information. Keep primary evidence linked to each research note.
Try this: Turn a paper-reading session into a structured evidence table.
Explore Notion AI Create
Descript An audio and video editing workspace for turning explanations into shareable educational material.
Try this: Edit a two-minute explanation of a paper and include an accessible transcript.
Explore Descript No matches. Try another topic or choose “All”.