Overview Version History Q & A Rating & Review
Codex Quota Stats
Codex Quota Stats shows Codex usage directly in the VS Code status bar. It
combines local token history from your Codex state with the current 5-hour and
7-day quota windows so you can see both token volume and quota pressure without
leaving the editor.
Repository: https://github.com/eddie-givens/codex-quota-stats
What It Shows
compact token counts such as 42M or 22.2K
5-hour and 7-day quota percentages in the status bar
quiet background refreshes that do not flash the status bar on every update
green, yellow, and red usage states based on current quota and context pressure
a hover tooltip with:
current context health
5-hour and 7-day quota bars with reset countdowns
a link to open detailed numbers
a detailed numbers panel with:
current thread title and model
latest completed response token usage
rolling local token totals for 5 hours and 7 days
quota reset times
additional model-specific limits when available
How It Works
The extension reads local Codex usage from your Codex home directory, usually:
~/.codex
When auth.json is present, it also requests quota data from:
https://chatgpt.com/backend-api/wham/usage
Requirements
VS Code 1.74.0 or newer
Python available as python, or a custom Python command configured in settings
local Codex state on the machine you are using
Installation
Install from the VS Code Marketplace, or install from a .vsix file:
Open VS Code.
Open the Command Palette.
Run Extensions: Install from VSIX...
Select the .vsix file.
Reload VS Code if prompted.
Commands
Codex Quota Stats: Refresh
Codex Quota Stats: Show Detailed Numbers
Codex Quota Stats: Open Codex Folder
Settings
localCodexStats.updateInterval
Refresh interval in seconds.
localCodexStats.pythonCommand
Python command used to run the bundled usage reader.
localCodexStats.codexHome
Optional override for the Codex home directory.
localCodexStats.accountId
Optional ChatGPT account ID to send with quota requests. Leave blank to use
the active account in ~/.codex/auth.json.
Privacy
local usage data is read from files on your machine
quota data is requested directly from OpenAI using your existing local Codex auth state
the extension does not require a separate API key
Troubleshooting
If local token data does not appear, confirm that Codex has been used on that machine.
If quota percentages do not appear, confirm that ~/.codex/auth.json exists and is current.
If quota data comes from the wrong ChatGPT account, set localCodexStats.accountId
to the desired account ID and refresh.
If Python is only available as python3, set localCodexStats.pythonCommand accordingly.
If the extension will not install, confirm that your VS Code version meets the minimum requirement.