Codex Usage Tray

Your Codex capacity and activity, one click away.

A Windows tray companion for usage limits, reset timing, and local session analytics.

Download · Getting started · Themes · Build from source

At a glance

Limits	Usage

Interface previews use illustrative sample data.

Keep live capacity close without another taskbar window. The limits view shows what remains; the usage view explains your local activity over time.

• Capacity at a glance. Five-hour and weekly meters, reset countdowns, and a two-bar tray icon.
• Readable summaries. Total tokens, calls, sessions, cached input, and average daily usage.
• Useful history. Switch between 1-day, 7-day, and 30-day views, with reset markers on the limits chart.
• Model detail. Compare token share and estimated API cost, with page controls for additional models.
• Personal appearance. Seven themes with coordinated surfaces, text, controls, and chart accents.
• Everyday convenience. Refresh every minute or every five minutes, and optionally start with Windows.

Getting started

1. Download the setup executable from the latest release.
2. Run the installer. It installs for the current user, creates a Start Menu shortcut, and launches the app.
3. Click the tray icon to open your limits. Choose View usage for session analytics.
4. Right-click the tray icon to choose a theme, adjust settings, or exit.

The installer uses %LOCALAPPDATA%\CodexUsageTray and replaces an older running copy. If Windows hides the icon, open the tray overflow to find it.

Requirements

• Windows 10 or later.
• An authenticated Codex CLI session with local authentication data.
• .NET 8 Desktop Runtime for framework-dependent builds. Self-contained builds include the runtime.

Reading the interface

Element	Meaning
Top tray bar	Weekly capacity remaining
Bottom tray bar	Five-hour capacity remaining
Red tray bars	The reset monitor reports a reset probability above 70%
Limits history	Remaining weekly capacity over the selected period
Cumulative token chart	Running token total across the selected calendar period
Estimated API cost	Approximate model usage cost, not a subscription charge or invoice

Click the range and refresh controls to cycle their values. Press Esc to close either popout. Both windows scale to your display without scrollbars; use Previous and Next to browse model details.

Themes

Select a theme from the tray icon’s right-click menu. Your choice is saved and applied to both windows.

Theme	Appearance
Default	Slate surfaces, soft text, and blue accents
AMOLED	Pure-black background with vivid accents
Dracula	Purple-gray surfaces with lavender highlights
Nord	Cool blue-gray surfaces and muted Nordic colours
Solarized Dark	Deep teal surfaces with blue and gold accents
Light	Bright surfaces with dark text
Glass	Subtly translucent dark-blue surfaces

View theme previews

AMOLED	Dracula	Light

Data and privacy

Session analytics are calculated locally. The app does not upload local session history or collect telemetry. It uses the existing Codex authentication token to request account usage; no separate sign-in form is required.

Source	Purpose
%USERPROFILE%\.codex\auth.json	Existing authentication for the account usage request
%USERPROFILE%\.codex\sessions	Local token counts, model activity, calls, and sessions
%LOCALAPPDATA%\CodexUsageTray	Local history, settings, caches, and diagnostic logs
ChatGPT usage service	Live account limits via /backend-api/wham/usage
Codex Reset Today API	Public reset announcements and probability estimates
Public model-pricing pages	Periodic pricing updates for cost estimates
GitHub	Release checks and update links

Reset likelihood comes from the public Reset Beacon forecast API, which publishes separate 24-hour and 48-hour probabilities with a freshness window. Historical reset markers come from the public Codex Reset Today API and are cached locally. Model pricing is refreshed from each observed model's official pricing page, including newly released and dated snapshot models, with built-in fallback rates and a local cache for offline use. Unavailable model rates are labelled in the interface. GPT-6 Astra includes fallback rates of $10 input, $1 cached input, and $50 output per million tokens, verified against the official model pricing. Estimates use standard token rates and do not account for service-tier multipliers, long-context surcharges, or cache-write pricing.

Build from source

Install the .NET 8 SDK or a compatible newer SDK on Windows, then run from the repository directory:

dotnet build -c Release

Publish a framework-dependent Windows x64 build:

dotnet publish -c Release -r win-x64 --self-contained false

Or publish a self-contained single-file executable:

dotnet publish -c Release -r win-x64 --self-contained true /p:PublishSingleFile=true

The published executable is written to:

bin\Release\net8.0-windows\win-x64\publish\CodexUsageTray.exe

Troubleshooting

• No account limits: confirm that Codex is signed in and the local authentication file exists.
• No usage history: analytics appear after Codex writes session activity. The limits chart needs a few refreshes to accumulate history.
• Unavailable cost: a matching model rate may not be available. Estimates can differ from actual billing.
• Startup or refresh problems: inspect %LOCALAPPDATA%\CodexUsageTray\crash.log if present. Remove sensitive information before sharing logs in an issue.

License and acknowledgements

Released under the MIT License. Reset likelihood is provided by Reset Beacon; historical reset markers are provided by Codex Reset Today.

This independent project is not affiliated with or endorsed by OpenAI. Official account usage remains authoritative; reset probabilities and API cost estimates are not guarantees or billing data.