Feature Request: Expose Claude.ai plan usage in status line

Resolved 💬 2 comments Opened Feb 1, 2026 by palonsorodriguez Closed Feb 1, 2026

Summary

When using Claude Code with a Claude Pro/Max subscription, there's no way to monitor plan usage limits without visiting https://claude.ai/settings/usage in a browser. This leads to unexpected rate limiting mid-session.

Feature Request

Expose Claude.ai plan usage data (percentage used and reset time) in the JSON payload passed to custom statusLine hooks, so it can be displayed alongside context window info.

Custom Status Line Example

Using a custom statusLine hook in ~/.claude/settings.json:

{
  "statusLine": {
    "type": "command", 
    "command": "node \"/path/to/statusline.js\""
  }
}

The status line could display:

Opus 4.5 │ my-project ███████░░░ 72% │ ██████░░░░ 64% ⏱ 2h15m
          ↑ model      ↑ context window  ↑ plan usage  ↑ resets in

Proposed JSON Payload

Currently, custom statusLine commands receive JSON with context_window. Please add plan_usage:

{
  "context_window": { "used": 72, "total": 100 },
  "plan_usage": { "used_percent": 64, "resets_in": "2h15m" }
}

Why This Matters

  1. Prevents surprise rate limiting - Users can pace their work knowing how much quota remains
  2. Better session planning - Know whether to start a complex task or wait for reset
  3. Enables custom displays - Users can build their own status line visualizations

Scope

This applies to subscription-based accounts only (Claude Pro/Max). API accounts (pay-as-you-go) have no usage cap.

Current Workaround Attempts

We attempted to scrape the usage page automatically, but Cloudflare's bot protection blocks:

  • Direct API/curl requests
  • Headless browsers (Playwright/Puppeteer)
  • Browsers with anti-detection flags
  • Even visible automated browsers get stuck in verification loops

This data can only be reliably accessed if Anthropic exposes it in the statusLine JSON payload.

---

Submitted via Claude Code session

View original on GitHub ↗

This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗