AWS introduces ToolSimulator, an LLM-powered framework within Strands Evals that enables safe, scalable testing of AI agents without risking live API calls or data exposure.

ToolSimulator uses LLM-powered simulations instead of live APIs to test AI agents, eliminating risks of exposing personally identifiable information (PII) or triggering unintended actions

Available now as part of the Strands Evals Software Development Kit (SDK), it replaces brittle static mocks with dynamic simulations that work across multi-turn workflows

Helps developers catch integration bugs early, comprehensively test edge cases, and deploy production-ready AI agents with greater confidence