Sign in
Enter a join code: run savetokens pair on a machine connected to this hub, or use the code
from the line savetokens server --setup printed.
Finish setting up
Or run savetokens setup: it finds your plans and asks for the rest, one question at a time.
Add a machine laptops, servers, CI, GPU boxes
Paste this on each machine. It installs savetokens if needed and joins this hub. It works for 24 hours, for up to 10 machines.
…For CI, containers and short-lived machines, set two environment variables instead:
Shown once. Keep it in your CI's secret store; anyone with it can send usage to this hub.
Spend so far, and where it's heading
Coming up
By provider
| Provider | Tokens | Spend | Likely by end | Next 30 days | $ / M tokens |
|---|
Limits
Usage per hour
Sessions, last 24 hours
| Session | Tool | Today | Last hour | If you stop it |
|---|
Today and last hour are % of each session's own limit: its plan's weekly limit, or its API budget.
Models this week
Machines this week
Recent alerts
Limits you've hit
Read-only. Only token counts, costs and limit readings reach this page: never prompts, replies, code or file contents.