Observe, route, and understand every model call.

Live tracing, per-model routing, and usage analytics for every LLM in your stack.