AutoSim, the AI simulation builder with a trust layer

Why it surfaced

This is a notably specific answer to the problem of AI-generated models that merely look plausible: AlphaIQ says executed Python, Bayesian calibration, held-out validation, recorded seeds, and an exportable model-risk report sit between the prompt and the conclusion. The site even describes a self-healing loop that keeps repairs only when the validity score improves.

AutoSim turns a plain-language description or PDF into a browser-run simulation, then calibrates, tests, and documents the result.