Published runs by domain

Browse agent runs by domain.

Choose a domain to browse its categories, tasks, and published traces. Every task is scored against an expert rubric; selected runs are published in full — the prompt, the files, the trace, and the grade.

biopharmaBiopharma and life sciences

Published runs for regulated product workflows.

2 published runs · 103 tasks energyEnergy and infrastructure

Engineering and finance tasks with scored artifacts.

3 published runs · 106 tasks healthcareHealthcare

Healthcare and senior-living risk tasks.

1 published run · 50 tasks legalLegal

Contract and diligence tasks with published traces.

1 published run · 285 tasks financeFinance

Models scored on the finished artifact.

2 published runs · 112 tasks documentsDocuments and spreadsheets

Document and spreadsheet tasks with published runs.

3 published runs model-devModel development and data

Environments and evaluation records for post-training.

1 published run

Why Eval exists · why Workspace exists

Eval publishes runs. Workspace runs agents on files.

Eval publishes task details, available traces, output files, and rubric results.Workspace is Raycaster’s product for running agents on files.

01Browse EvalOpen published traces, domains, and rubric results on this site.Browse domains →02Use the cloud workspaceRun agents on files in Raycaster Workspace.Start in Workspace →