claude-code - 💡(How to fix) Fix [FEATURE] Enterprise usage bar support [1 participants]

Preflight Checklist

• I have searched existing requests and this feature hasn't been requested yet
• This is a single feature request (not multiple features)

Problem Statement

<img width="476" height="110" alt="Image" src="https://github.com/user-attachments/assets/5f9f8154-52a5-4ee3-89bb-e54334306039" />

Claude Code displays a Usage bar in its status line, but enterprise account users see -- instead of their actual spend data. This is because Claude Code fetches usage via the Anthropic API key, while enterprise usage limits are tracked through the claude.ai web session (SSO/cookie auth). The two auth paths don't share data, so Claude Code has no way to surface the enterprise spend limit.

The data is available — the Claude desktop app shows it correctly in Settings → Usage (e.g. "$88.57 of $150.00 · 59% · Resets May 1"). <img width="943" height="195" alt="Image" src="https://github.com/user-attachments/assets/3ec80c9a-aa8b-467d-87e3-b840963e937a" />

Enterprise users are left with no way to monitor their spend from within their primary dev workflow, and must context-switch to the desktop app to check it manually.

Proposed Solution

1. Expose an enterprise usage endpoint accessible via API key. Add a /v1/usage (or /v1/organizations/{id}/usage) endpoint that returns current spend, limit, and reset date for enterprise accounts, authenticatable with the same API key Claude Code already uses. Claude Code can then poll it and render the usage bar identically to how it works for non-enterprise accounts.

Alternative Solutions

1. Let users configure a session token for usage fetching. Add an optional config field (e.g. usageSessionToken) that Claude Code can use to query the claude.ai usage API separately from the API key used for inference. This keeps inference auth unchanged while allowing usage data to flow through.

2. Surface the data already available in the desktop app. If the desktop app already caches or fetches enterprise usage locally, expose it via a local IPC or file so Claude Code (running on the same machine) can read it without any new network auth.

Any of these would close the gap. Option 1 proposed above seems like the cleanest long-term fix.

Priority

High - Significant impact on productivity

Feature Category

CLI commands and flags

Use Case Example

I'm working on something in Claude Code CLI and I can see where my current Enterprise usage is, rather than guessing, or having to cross check and context switch with the desktop or web app, open settings -> usage page.

Additional Context

No response