Work through a complete example, inspect the handoff between steps, and keep a record you can revisit. Each tutorial supplies small inputs and a defined stopping point.
Choose a walkthrough
Fold a 76-residue protein, inspect confidence, and compare with an experimental structure.
Start the walkthrough → 02 / Two stagesPrepare a receptor and dockSeparate a public complex, repair its receptor, and review predicted ligand poses against a reference.
Start the walkthrough → 03 / Three inputsA bounded bulk workflowMove from a single pilot to three independent predictions while preserving candidate identities.
Start the walkthrough →Before running a tutorial
Use a signed-in account with access to the selected services. Review the current form and credit estimate at each submission. Downloads and local preparation do not submit jobs; running the hosted steps consumes compute credits.
These are worked procedures, with explicit inputs, settings, and review checkpoints. They do not contain precomputed Neurosnap predictions or promised scores. Runtime, filenames, and results depend on the service version and settings; use the actual job output listing.
Keep a research record
Record the input identifier, exact sequence or file, tool version, settings, job ID, output filenames, and your interpretation. The three-input run manifest provides a starting point. Keep experimental references separate from predictions and label synthetic variants clearly.
For method selection, use the scientific task guides. For custom orchestration, continue with the API quickstart or Python SDK.