[Bug] claude.ai and Claude Code show different usage data for the same account

Preflight Checklist

• [x] I have searched existing issues and this hasn't been reported yet
• [x] This is a single bug report (please file separate reports for different bugs)
• [x] I am using the latest version of Claude Code

What's Wrong?

What I encountered:

• claude.ai/settings/usage shows remaining quota
• Claude Code shows "limit reached"
• Same account, two different results
• After switching to my second Pro account on the same device, Claude Code burns through tokens abnormally fast
• Clearing all local data (~/.claude/, keychain, sessions, projects, statsig) does not fix it — the problem is server-side

My observation:

Anthropic recently announced that 70-90% of company code is AI-generated, and Claude Code itself is ~90% written by Claude Code.

100% AI coding is fine. But it doesn't mean you can ignore system architecture.

Two clients (claude.ai and Claude Code) querying usage data for the same account should return the same result. They don't. This is a clear violation of the Single Source of Truth principle.

Can system architecture really allow such a fundamental principle to be violated?

What Should Happen?

Two clients (claude.ai and Claude Code) querying usage data for the same account should return the same result.

Error Messages/Logs

Steps to Reproduce

See description above.

Claude Model

Opus

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

v2.1.5

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_