FAQ

FAQ

Is it safe?

It reads another app's keychain item and calls undocumented endpoints, so the questions are fair.

• Open source (MIT): every line is in the repo.
• The only network calls are to the providers' usage endpoints and your own ntfy server, if configured. No analytics, no telemetry.
• It never refreshes a local account's token. See How it works.
• Remote credentials are stored in a device-only keychain item, never in UserDefaults, never logged.
• Releases are self-signed with a stable identity, so permissions survive updates (not notarized, so first launch still needs Open Anyway); the app is not sandboxed. See Installation.

Which providers are supported?

Claude Code (local and remote accounts) and Codex (local ChatGPT-login accounts). See Accounts.

Why is there no Codex spend row?

The Codex payload reports a remaining credit balance, not an amount spent, so there is nothing honest to show.

Why does the menu bar label have no color?

SwiftUI renders it as a template image, so it cannot carry severity color. Color lives in the popover and toasts.

Why is the weekly countdown approximate?

Weekly resets_at is the server's own rolling-window estimate.

Can I poll faster than every 180 seconds?

Settings offers shorter intervals, but local Claude accounts share a usage cache that reuses a response younger than 180 seconds, and the endpoint rate-limits aggressively. See How it works.

Will it alert on every poll?

No. Alerts fire only when a level rises within a reset window. See Alerts.

Does it start at login?

Optional: Settings → General → Start at login.

Where do I report a bug?

Issues. Include relevant log lines from Troubleshooting.