jev-auto-router

An experimental prototype (documented in Chinese) that has Jev pick which GPT model and reasoning tier should handle each individual call within a Codex session, routed through a local proxy that keeps the session and tool loop continuous, with task correctness checked independently afterward to see whether using fewer flagship-model calls still gets the job done at lower cost.

unverified — sourced from awesomejev.com; description independently written by AionEdge from the repo’s own GitHub metadata and README, not GitHub-API relevance-checked