Skip to content

API reference

Area APIs
Tasks and runs Queries, task families, predictions, datasets, and serialized runs
Benchmarks Loader protocol, dataset registry, task selection, installation, and runtimes
Agents and tools Strategy protocols, configurations, user simulation, tools, and tracing
Pipelines Prediction, execution, evaluation, and ensembling
Preprocessing Derived schemas, summaries, embeddings, and ER diagrams
Metrics Metric protocol, registered metrics, and aggregation policies

See Extend the toolkit for worked examples and the library reference for connectors, schemas, execution results, and model infrastructure.