Evidence level

A label for how trustworthy a reuse measurement is, from provider_reported down to trace_estimated and unknown.

Not all reuse numbers are equal. provider_reported comes from the provider’s own cached-token usage; runtime_confirmed comes from a BYOC runtime; router_inferred is a prediction; trace_estimated is a diagnostic estimate; unknown means the adapter has no trustworthy signal.

Zumik attaches an evidence level to every capture figure so a confident estimate is never mistaken for a measurement.

Keep reading.

Cache capture rate

Realized reused tokens divided by candidate reusable tokens - how much of the available reuse a provider actually delivered.

Reuse opportunity

The maximum share of input tokens that could be served from cache, independent of whether they actually were.

See it in practice.

Definitions are useful; measurement is better. Run a diagnostic on your own workload.

Run a diagnostic Back to glossary