Hi, I'm Aldrin, the maker of TokenHub.

I use several AI coding agents across different subscriptions and wanted one place to understand their recorded token usage.

TokenHub runs locally and organizes that usage by agent, model, and session. A lot of the work has been in reconciling different formats, avoiding duplicate counts when history is copied or imported again, and making missing counters visible.

It doesn't need provider API keys. You approve the local sources it imports, and it retains usage metadata rather than transcript content. Cost estimates use editable rates and should not be read as invoices or subscription balances.

Install with `pip install tokenhub`, then run `tokenhub` to open the dashboard.

The gallery shows the TokenHub 0.1.6 interface with generated sample data.

I'd love feedback on the setup experience and which usage breakdowns would help you most.