Gaugelet privacy notice

Last updated: August 21, 2026

Summary

Gaugelet is a local-first menu-bar app. It has no Gaugelet account, advertising, analytics, telemetry, crash-reporting service, usage-history upload, or LammWorks backend.

Gaugelet communicates with two external boundaries:

1. The separately installed Codex CLI may contact OpenAI when Gaugelet asks its local App Server for rate-limit data.
2. Sparkle may contact GitHub to check for and download Gaugelet updates.

Gaugelet does not sell personal information.

Codex usage data

For live mode, Gaugelet starts the locally installed codex app-server, initializes a private standard-input/output session, and sends a read-only account/rateLimits/read request.

The response may contain:

• Usage-window names or identifiers.
• Percent used or remaining.
• Window duration and reset time, when returned.
• Plan or model labels, when returned.
• Rate-limit, workspace-credit, or spending-control state, when returned.

Gaugelet does not ask this integration for conversation content, prompts, responses, message history, project files, or billing transactions.

Authentication boundary

The Codex CLI owns authentication. Gaugelet does not read or copy browser cookies, browser storage, Codex authentication files, access or refresh tokens, API keys, or passwords.

The Codex CLI may use its existing session to communicate with OpenAI. OpenAI and the Codex CLI process data under their own terms and privacy notices; Gaugelet cannot control that separate relationship. Gaugelet does not proxy Codex traffic through LammWorks.

Update checks

Gaugelet includes Sparkle 2.9.5. If you accept automatic update checks, Sparkle normally requests this GitHub-hosted appcast about once every 24 hours. A manual Check for Updates request uses the same channel:

https://github.com/lammworks/Gaugelet/releases/latest/download/appcast.xml

If you approve an update, Sparkle downloads the selected DMG from GitHub Releases. GitHub and its network providers may receive ordinary network metadata such as the source IP address, request time, requested URL, transport details, and user-agent information. Their handling of that data is governed by their own policies.

Gaugelet disables Sparkle system profiling. It does not deliberately attach a hardware or macOS system profile to update requests. Gaugelet also disables silent update installation: download and installation require user confirmation.

Sparkle verifies the signed update feed and enclosure before extraction. Update authentication does not make the initial Gaugelet download Apple-notarized.

Local storage

Gaugelet stores preferences with the standard macOS preferences system, including settings such as:

• Menu-bar percentage and hover behavior.
• Warning threshold and notification preference.
• Launch-at-login intent and status-related state.
• Selected icon theme.
• Selected source mode and demo scenario.
• Sparkle update-check preference, managed by Sparkle.

Live percentages and reset timestamps remain in process memory. A last successful reading may remain in memory briefly enough to be shown as explicitly stale after a refresh failure. Gaugelet does not maintain a persistent usage-history database.

When notifications are enabled, Gaugelet asks macOS to deliver a local notification when a live window crosses the selected threshold or reaches its limit. Demo and stale readings never trigger an alert. Gaugelet does not send notification content to LammWorks.

macOS and GitHub may independently retain operational records such as system logs, crash logs, quarantine data, access logs, or release-download logs. Gaugelet does not upload macOS logs.

Demo, stale, unavailable, and signed-out states

Demo data is generated locally and labeled as demo. If a refresh fails after a successful reading, Gaugelet may show the last reading as stale with its original update time and the failure reason. If no successful reading exists, Gaugelet shows unavailable or signed out. It does not replace missing account data with plausible percentages.

Retention and deletion

Because Gaugelet operates no backend, there is no Gaugelet server account or server-side usage history to delete.

Removing Gaugelet deletes the app but may leave its macOS preferences. To remove those preferences, quit Gaugelet and run:

defaults delete com.lammworks.gaugelet

This affects Gaugelet preferences only. It does not modify Codex CLI credentials or OpenAI account data. Use Codex or OpenAI controls for those separate systems.

Information you choose to share

GitHub processes information you voluntarily submit in an issue, pull request, security advisory, or other repository interaction. Remove account names, file paths, tokens, cookies, keys, prompts, source code, and other sensitive material before sharing diagnostics.

Use GitHub Issues for non-sensitive privacy questions. Use GitHub Private Vulnerability Reporting for a suspected security or privacy vulnerability. No support email is represented by this project.

Changes

This notice must be reviewed before Gaugelet adds analytics, crash reporting, persistent usage history, another data provider, account functionality, a LammWorks service, or materially different update behavior. Material changes are recorded in CHANGELOG.md.

Gaugelet is independent software and is not affiliated with, endorsed by, or sponsored by OpenAI.

Built by LammWorks.