BioSTEAM / QSDsan MCP quickstart

Build or load a process system, inspect it, and keep the result identity consistent across analyses.

Start from a published model

text

health_check
→ list_biorefinery_models
→ build_system(model_name=...)
→ get_stream_results(result_id=..., stream_id=...)
→ get_unit_results(result_id=..., unit_id=...)
→ get_tea_results(result_id=...)
→ dispose_result(result_id)

Build a custom process

Use build_system_from_spec when you need a small structured flowsheet instead of a published biorefinery. Keep the chemicals, thermodynamic assumptions, unit parameters, stream specifications, and source revision with the study.

Uncertainty and optimization

Use run_uncertainty and run_sensitivity only after the reference process is stable. Use optimize only with explicit variables, bounds, objectives, and engineering constraints.

For long operations, run_uncertainty_async and optimize_async return a job_id that can be managed with the job-lifecycle tools.

Stored process identity

Build calls return a result_id. Version 0.4.1 can restore reconstructible process-system handles when persistence is enabled. The persisted record is a build recipe and runtime fingerprint, not a serialized scientific Python object.