Local Inference Leakage Tracker

Coins (tokens) and dollars saved by keeping inference on the local GPU instead of the paid cloud.
Loading…
Coins kept local (all-time)
 
Dollars saved (all-time)
 
Leakage today
 
Cloud spend this week
 

Where the tokens went

Today
This week
Kept local (free)Leaked to cloud (paid)

Local vs cloud tokens per day

By model

ModelWhereTokensRequests
Data source: Quinn usage meter. Local inference = Ollama on the ALIEN GPU box. Savings value each local token at your realized cloud rate (this week's Claude cost ÷ Claude tokens). Use the Reload button in the header to pull fresh numbers.