Review requests recorded in this workspace, including measured latency, token usage and available cost data.
No provider-reported successful costs available yet.
Projection uses the average of 0 successful requests with returned cost, across your local model mix. Unknown costs, failures, future price changes and changes in prompt length are excluded. This is not a spending limit.
Your first request will begin a personal latency timeline.
Execute a request with your own provider key. History is in memory unless you enable device storage.