Analytics and visibility

Usage analytics show request volume, token usage, model usage, source split, and spend estimates so teams can understand how AI is being used across web and API clients.

Routing analytics show Direct versus Flux usage, provider/model load, fallback pressure, latency, reliability, and route outcomes. This helps identify whether Auto mode is making useful decisions.

Knowledge analytics show memory health, graph coverage, truth status, evidence, citations, stale items, and the queue of workspace knowledge that needs review.