FAQ
The windows, the meters, and the caveats.
Tokenomics shows each provider's usage in that provider's own units — and providers don't agree on how to count. This page explains what the numbers mean, why some tabs show two meters and others one, and how sign-in and privacy work.
Reading your usage numbers
How do Claude's 5-Hour and 7-Day windows work?
These are Anthropic's own quota windows, not something Tokenomics invented. The 5-Hour Window is the rolling cap most people hit first — it resets about five hours after your first message in the current block. If Tokenomics shows 80% with two hours left, you're close to a rate limit, so pace your prompts or switch tasks.
The 7-Day Window is the longer ceiling, resetting weekly from your first message of the cycle. On Pro plans you'll also see an Opus sub-counter inside the 7-day window — Opus carries its own weekly cap on top of the shared total, so it can run dry while the 7-day window still has room.
Why does Anthropic show only one window when other providers show two?
Anthropic is the exception. Claude on the web, the Claude desktop app, and Claude Code all draw down the same 5-hour and 7-day windows on your subscription — one set of limits covers every surface, so the Anthropic tab needs only the 5-hour and 7-day split.
OpenAI and Google structure their products the other way: the same account has a separate meter per surface. That's why those brands render two bars instead of one — see the next answer.
Why does my ChatGPT tab show two progress bars?
ChatGPT and Codex CLI are two different products that share one OpenAI login but bill against separate limits. Consumer ChatGPT runs on its own 3-hour and weekly windows; Codex CLI has its own weekly quota tied to your subscription tier. Same account, different meters — and OpenAI doesn't return a combined view, so Tokenomics renders them as two bars to show which limit you're closer to.
The same is true for Gemini: gemini.google.com and the Gemini CLI share a Google account but track usage independently.
Why does my widget show OpenAI twice?
Widgets render each usage pool as its own row, because the compact format has no room to nest. If you've connected both ChatGPT and Codex CLI, the OpenAI brand contributes two rows. On the small widget you pick exactly one pool to show — open the widget's edit panel and choose ChatGPT or Codex CLI.
Upgrading from an earlier version? If your small widget was set to "OpenAI" or "Google AI," it now maps to Codex CLI or Gemini CLI respectively — those entries were already tracking the CLI pools, since chat readers for those brands didn't exist before. To show the chat pool instead, re-select it from the edit panel.
Accounts & privacy
Is my usage data sent anywhere?
No. Tokenomics reads usage from the AI tools you already use, holds the numbers in memory while it's open, and sends nothing to any server the author controls — no analytics, no telemetry, no third-party SDKs inside the app. The full, auditable breakdown is on the Privacy page.
How does signing in work?
For every provider that ships a command-line tool, Tokenomics reads the credentials those tools already stored on your Mac — you never hand Tokenomics a password, and there's nothing to paste. The few providers without a CLI use an API key you paste once, kept in the macOS Keychain. Tokenomics never extracts or copies your tokens anywhere else. The specifics are itemized on the Privacy page.
How do I uninstall Tokenomics?
Quit the app, then drag Tokenomics from your Applications folder to the Trash. Tokenomics keeps nothing of its own on disk while it runs. The only possible leftover is the update checker's saved preference; for a completely clean removal you can delete Tokenomics' entry from ~/Library/Preferences, but it's harmless to leave.
Still stuck, or hit something this page doesn't cover? Email [email protected] — or open an issue on GitHub.