Skip to content

Loop Engineering

Loop Engineering is an alpha Python runtime for building repeatable agent workflows with explicit state, budgets, deterministic checks, recovery, checkpoints, and persistence.

Start here

git clone https://github.com/chillum-codeX/loop-engineering.git
cd loop-engineering
pip install -e .
loop-engine --help

Continue with the five-minute quickstart, then work through the tutorial series to understand the runtime itself, or jump straight to a starter pattern.

Evidence level

  • 146 automated tests pass locally and in GitHub Actions.
  • Deterministic benchmark controls validate evaluator behavior, not model superiority.
  • One paid OpenRouter smoke request verified provider-reported token and cost accounting without storing its prompt, response, or credential.
  • Historical mock benchmark outputs are not model-performance or security evidence.

Loop Engineering makes no SOTA, production-readiness, or comparative performance claim.

Current release

Loop Engineering v0.4.1 is the current verified alpha release.