Model
Select and run computational tools, then inspect their outputs and provenance.
Model
The Model menu contains Tool Selector, where you choose computational tools from RxCubed and neurosnap.ai.
Select and run
Search or filter tools by provider, modality, and functional category. Select a tool to open its form. File inputs are chosen from compatible evidence already saved in the project, which preserves provenance.
Enter an editable run name and settings, then click Run. A run is registered immediately. Fast tools finish in the request; queued and remote jobs report progress in Run Results. A failed run writes no outputs.
Providers
RxCubed tools run on the platform with no external key. neurosnap.ai tools use your configured provider key and may incur provider charges. The tool page shows the provider and required settings.
Results
Computational outputs are saved automatically as evidence; there is no second save step. Each output belongs to its run and points to the selected input evidence. All Runs is the computational slice of Results Hub and uses the same navigator and evidence cards.
The Results Hub hierarchy is Category → Tool → Run → Evidence. A computational run's outputs sit under that run; a collection run holds the file types you selected for one search result.
Related documentation
Filling in a tool and collecting its results covers file and text inputs, checking a file with File info ▸, finding a file you don't have yet, and collecting asynchronous runs from Run Results.
Use Tool Selector, RxCubed Tools, and the individual tool pages for input requirements and interpretation.