Show HN: TrainForgeTester – deterministic scenario tests for AI agents

Category: devtools

Tags: ai-testing, agent-testing, regression-testing

Score: 7.3/10 (Innovation: 7, Technical: 7, Documentation: 8, Utility: 7)

TrainForgeTester provides a deterministic-first approach to regression testing of AI agents, using golden injection and Python equality for tool calls and exact text, with LLM-based binary yes/no checks only for natural-language divergence. Its innovative combination of golden injection, deterministic checks, and stable binary LLM judging addresses a known gap in reliable agent testing.

Target audience: backend devs, ai engineers, data engineers

Repository: https://github.com/TrainForge/TrainForgeTester · Python · NOASSERTION · 2 stars

View on Hacker News