LiteLLM

LiteLLM has become the de facto standard LLM proxy layer, used by Netflix and other large-scale adopters to unify access across providers. Its budget-aware routing can automatically switch to cheaper models when spend thresholds are reached, enforcing Token Budget limits at the per-key, per-user, and per-team level. This makes it a natural fit for agentic systems where multiple agents consume tokens independently and cost governance requires granular controls. The MIT-licensed core is free to self-host, with an Enterprise tier at $30K per year for advanced features.