🏠 Keep your AI at home — local-first tiered LLM routing. ~90% of traffic served by open-weight models on your own hardware, ~10% budget-capped frontier (Claude/GPT/Gemini), a privacy-pinned lane that can never leave your box, a pluggable gateway (LiteLLM · Bifrost · Helicone), and Grafana dashboards for all of it.

Automatically use the best of Claude Sonnet 3.5, OpenAI o1, GPT-4o or Gemini in a single LLM call with ChatLLM's RouteLLM.

An intelligent LLM routing system optimizing the cost-quality trade-off for RAG pipelines. Uses multi-tier semantic complexity scoring and embedding models to dynamically route queries.

Cost-aware LLM router on the cost-quality frontier. Benchmarked on current frontier models (LLMRouterBench: GPT-5, Claude-4, Gemini-2.5-Pro, Qwen3, DeepSeek): matches the top model's quality at 98% lower cost, dominates all 12. Model-agnostic: bring any pool.

A Streamlit chat app that uses RouteLLM to route prompts between GPT-4o-mini and Nebius Llama.