Claude, Codex, Cursor, Gemini and Copilot side by side, with a countdown to every reset,
so the next "usage limit reached" never catches you mid-flow.

Why AI-Usage

You are deep in a refactor, the agent is on a roll, and then the limit hits. Which tool still has room? When does the 5-hour window reset? How close is the weekly cap?

AI-Usage answers all of that at a glance. It sits in a corner of your screen (or just in the tray), reads the numbers straight from the tools you already use, and turns them into clear bars, countdowns and history charts. No accounts to create, no API keys to paste, nothing to configure before the first number shows up.

Supported agents

Agent	What you see	Where the numbers come from
Claude Claude Code	5-hour and weekly limits, every extra window the account reports, token statistics	The sign-in Claude Code already has on this PC and its local session logs; an optional browser session for further accounts
OpenAI Codex	5-hour and weekly limits, real token counts, token statistics	Codex's local session files; a signed-in browser session as a fallback
Cursor	Monthly plan usage, split into Cursor's own models and API models where the plan reports it	A signed-in browser session
Google Gemini Antigravity	Gemini quota windows with their reset times	The account the Antigravity CLI (agy) is signed in with; a signed-in browser session for the extra figures
GitHub Copilot	Monthly premium requests	The GitHub CLI (gh) you are already signed in with

More than one account per agent? Add it, and it gets its own tile and history. Copilot offers every further account the GitHub CLI is signed in with; the other agents sign in through the built-in browser.

_{Every supported agent at once, small layout}

Highlights

At a glance

• Every usage window an agent reports, as a bar with its percentage and a plain countdown ("resets in 2h 14m"), plus its tokens and a "full in about 3h" forecast when there is room
• A warning sign next to a quota that runs low, so the level never depends on colour alone
• A tray icon that shows the busiest quota, or the one agent and window you pick
• Threshold alerts per agent and window, plus a notice once a window is free again
• History per agent over 24 hours, 7 days, 30 days, a year or everything, kept locally for up to ten years
• Error messages that name the cause (no answer in time, HTTP code, no connection) and offer to sign in again when a sign-in has expired

Made for your desktop

• Tiles below each other or side by side; the window sizes itself and switches between the full and the small layout so nothing is cut off
• Always on top, click-through overlay mode and adjustable opacity
• Themes: System (follows Windows light and dark live), Nebula, Terminal, Dark and Light, plus high contrast
• Autostart, timed refresh, settings export and import, update check
• Light on your PC: while the screen is locked the widget slows down and catches up once you are back; the hidden sign-in browser is released after 15 idle minutes
• Remove an account in one step: its sign-in in AI-Usage goes, and its history too if you tick it
• English and German

Token usage window

Quotas tell you how much is left. The token usage window shows where it went. It reads the session logs Claude Code and Codex keep on your PC, builds its own local index and turns them into a dashboard for any period you pick: 7 days, 30 days, a year or everything.

• Your history at a glance: an activity grid with one square per day, colored by the agent that led that day. It reaches back as far as the window is wide, at least twelve months. Click a day to open its details: agents, models, projects and hours.
• Key figures: total tokens, average per day and per active day, the busiest day and the change against the previous period.
• Breakdowns: input, output and cache tokens, grouped by day, week, model, project or effort level, plus the share per agent, model and effort level.
• Your projects: the folders that used the most tokens, with their share of the total.
• Patterns: usage per day, per weekday or per hour.
• CSV export for your own spreadsheets.
• Your layout: drag a section by its heading above, below or beside another one, or move it with Alt+Up and Alt+Down. The layout is remembered and can be reset.

Cursor and Copilot keep no token counts on your PC, so the window says so instead of showing a guess. Gemini counts come from the local conversations of the Antigravity CLI.

_{The token usage window}

Keyboard shortcuts

Where	Keys	What it does
Widget	F5	Refresh now
Widget	Ctrl+,	Open the settings
Widget	Ctrl+T	Open the token usage window
Widget	Ctrl+L	Open providers and layout
Widget	Alt+Space	Open the window menu
Anywhere	Ctrl+Alt+U	Show or hide the widget. This is the default; the shortcut is off until you turn it on in the settings, where you can also change the keys
Anywhere	Ctrl+Alt+Left, Ctrl+Alt+Right	Snap the widget to the left or right half of the screen
Anywhere	Ctrl+Alt+Up, Ctrl+Alt+Down	Snap the widget to the top or bottom half of the screen
Token usage window	Esc	Close the window (first cancels a running drag)
Token usage window	Alt+Up, Alt+Down	Move the focused section up or down
Token usage window	Arrow keys, Home, End	Move through the bars and days of a chart

Use it in your terminal

While AI-Usage runs, it keeps a small file up to date with the numbers it shows: %APPDATA%\AI-Usage\status.json (in the data folder you chose, if you moved it). It is written a couple of seconds after each reading, and always as a whole, so a program reading it never sees half a file. Hidden tiles are in it too.

{
  "version": 1,
  "updated": "2026-10-10T14:03:09+02:00",
  "providers": [
    {
      "id": "claude",
      "name": "Claude",
      "account": "Work",
      "status": "ok",
      "windows": [
        { "kind": "fiveHour", "label": "5 hours", "usedPercent": 42, "resetsAt": "2026-10-10T16:17:00+02:00", "tokens": 1823457 }
      ]
    }
  ]
}

• status is ok, stale, failed, signin (signed out or blocked) or unavailable (nothing to read on this PC). After a failed reading the last numbers stay in windows.
• kind is fiveHour, weekly or other; label is the text the widget shows, in its language. resetsAt and tokens are null when unknown.
• account is the name you gave the account in the settings, or null.

AI-Usage.exe --status prints one line per provider from that file and exits. It opens no window and works whether the widget is running or not; when the file is older than 15 minutes, each line says how old it is. --status --json prints the file itself. Without a status file the exit code is 1.

Claude (Work) 42% 5h (2h 14m) · 63% week (3d 4h)
Codex 7% 5h (45m)

AI-Usage is a window program, so in a plain console the lines can show up after the prompt is back. Pipe the output (AI-Usage.exe --status | more) or use it from a script, where it behaves as usual.

PowerShell:

Get-Content "$env:APPDATA\AI-Usage\status.json" | ConvertFrom-Json | ForEach-Object providers

Bash with jq:

jq -r '.providers[] | "\(.name) \(.windows[0].usedPercent)%"' "$APPDATA/AI-Usage/status.json"

Claude Code status line (use the full path of your copy):

{
  "statusLine": {
    "type": "command",
    "command": "C:/path/to/AI-Usage.exe --status"
  }
}

The file holds only numbers and the names you chose. It never contains an e-mail address or the account name a provider reports, and it stays on your PC.

Private by design

• Reads usage numbers only. AI-Usage only calls the usage pages and endpoints each agent offers. It never sends a prompt to a model, so reading your quota does not spend any of it. A test in the suite checks every endpoint the app can call.
• Leaves your tools alone. It never writes to an agent's own files and never copies or stores the sign-in of your CLI tools. Browser sign-ins live in AI-Usage's own browser profile per agent. Signing out in AI-Usage only disconnects AI-Usage; your CLI and IDE stay signed in.
• Project colors from your project icons. To give each project in the token usage window its own color, AI-Usage looks inside the project folders named in the session logs for an app icon (.ico or .png, up to six folder levels deep) and takes its main color. It only reads those images and keeps nothing but the color.
• Stays on your machine. Settings, history and the statistics index live in %APPDATA%\AI-Usage\. Browser sign-ins get their own profile per agent under %LOCALAPPDATA%\AI-Usage\webview\. Apart from the requests to each agent's usage pages, the Google token refresh for an expired Antigravity CLI sign-in, the update check and, only when you agree to it, the download of Microsoft's WebView2 installer, nothing leaves your computer. No telemetry.
• Speaks like your CLI. To read your limits, the Claude usage request identifies as Claude Code, and the Antigravity token refresh uses the client id and secret that ship inside your installed Antigravity CLI. Nothing else is sent, and nothing is read from the CLIs' files except what that request needs.

Get started

1. Download AI-Usage.exe (portable) or the installer from the latest release.
2. Run it. Agents you are already signed in to show up on their own.
3. For the rest, open Settings > Providers and sign in through the built-in browser.

File	What it is
AI-Usage.exe	Portable for x64, no installation
AI-Usage-arm64.exe	Portable for ARM64, no installation
Setup-AI-Usage-<version>.exe	Installer for x64 and ARM64, for everyone or just for you, with optional autostart and desktop shortcut
SHA256SUMS.txt	Checksums for every file
*.exe.sig	Detached signature next to each exe, checked by the in-app update

What changed in each version is listed in the changelog.

Requirements

• Windows 10 (version 1809, build 17763) or Windows 11, x64 or ARM64
• The .NET 10 Desktop Runtime. The installer sets it up for you; the portable exe shows a download link when it is missing.
• Browser sign-in uses the Microsoft Edge WebView2 runtime; if it is missing, the app offers to install it with Microsoft's signed installer and everything else keeps working
• Copilot needs the GitHub CLI (gh) signed in; Gemini reads the Antigravity CLI's (agy) active account

Updates

When a newer release exists, a notice bar offers to install it. The installed copy runs the new setup, the portable copy swaps its own exe and restarts. Updating is optional: switch the check off under Settings > System and the app never asks.

Uninstalling keeps your data. Delete the two folders named above to remove it.

Coming from version 1.1.0

Up to version 1.1.0 every download carried its own copy of .NET, which made it about 136 MB. From this version on AI-Usage uses the .NET 10 Desktop Runtime installed on your PC, and the downloads shrink to a fraction of that.

Installed copy: the update runs the new setup, which checks for the runtime and, if it is missing, downloads it from Microsoft and installs it. Windows asks once for permission.

Portable copy: the update swaps the exe as before. If the runtime is missing, Windows shows a message with a download link instead of starting AI-Usage. Install the .NET Desktop Runtime 10 for your PC (x64 or Arm64) from https://dotnet.microsoft.com/download/dotnet/10.0 and start AI-Usage again. Your settings and history stay where they are.

To check beforehand, run dotnet --list-runtimes in a terminal and look for Microsoft.WindowsDesktop.App 10.

Build from source

Requires the .NET 10 SDK and, for the installer, Inno Setup 6.

build.bat

build.bat compiles, runs the test suite, publishes x64 and ARM64 as single-file executables that use the installed .NET runtime, builds the installer and writes everything plus SHA256SUMS.txt into dist\. When a signing key is present (.signing\ai-usage-update.pem, never committed, or the path in AIUSAGE_SIGNING_KEY) it also signs each exe; without one the build says so and the in-app update refuses that build.

Contributions are welcome, see CONTRIBUTING.md. Found a security issue? Please follow SECURITY.md.

FAQ

Another tool shows far more tokens than AI-Usage. Which one is right? Claude Code often writes the same response to its log several times. AI-Usage counts each response once; tools that add up every line count those repeats again, often close to twice the real total.

An agent shows no numbers. Open Settings > Providers and sign in there. Copilot needs the GitHub CLI signed in (gh auth login), Gemini reads the account of the Antigravity CLI (agy).

A tile says an agent is not answering or having problems. AI-Usage keeps the last numbers, dimmed, and tries again on its own. If it lasts, pick Check Claude status (or your agent's name) under the reason to see whether the agent has an outage.

Numbers update less often on battery. With Save energy on battery on, AI-Usage fetches half as often on battery and only reads local sources while Windows energy saver is on. Switch it off under Settings > Refresh.

My token history starts later than I expected. The token usage window can only read the session logs that still exist. Claude Code deletes its logs after 30 days by default; raise cleanupPeriodDays in its settings to keep more. Whatever AI-Usage has read once stays in its own index.

License

MIT. Third-party components: CommunityToolkit.Mvvm (MIT), Microsoft.Data.Sqlite (MIT), SQLitePCLRaw (Apache 2.0), Microsoft Edge WebView2 (Microsoft's own redistributable licence).

Claude, Codex, Cursor, Gemini, Antigravity and GitHub Copilot are trademarks of their respective owners. AI-Usage is an independent project and not affiliated with any of them.

_{© 2026 BGCoding}