Show HN: Levi – run AlphaEvolve on your Claude Code/Codex for dirt cheap

Category: ai-ml

Tags: evolutionary-search, llm-optimization, code-gen

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

LEVI is an LLM-guided evolutionary framework for code and prompt optimization that uses a behavioral archive and cost-aware model routing to achieve strong results with cheap local models, outperforming expensive frameworks like AlphaEvolve at a fraction of the cost. It is interesting because it decouples search quality from model size, demonstrating that better harness architecture can substitute for larger LLMs.

Target audience: backend devs, data engineers, ai researchers

Repository: https://ttanv.github.io/levi/ · Python · MIT · 22 stars

View on Hacker News