modelsteward v0.1.0 — Measured, not guessed: a llama.cpp server manager that tunes local models to their best performance and keeps OpenCode (and any OpenAI-compatible app) configured from real measur

Sign in to join the discussion — your thebay.events account works here.