Claude Code blocked by "weekly limit reached" (resets Jul 25) while all usage meters show 0% — claude.ai web works on 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?
On a Claude Max (20x) subscription, Claude Code refuses every request with:
You've hit your weekly limit · resets Jul 25, 9pm (America/New_York)
(banner also shows: "resets in 6d")
This occurs even on a single trivial prompt with no subagents, so it is not a
burst/concurrency issue. It affects the Claude Code CLI (VS Code integrated
terminal) and the Claude Desktop "Code"/Cowork surfaces, which the desktop
banner labels "shares limit with Claude Code".
Every visible usage meter shows the weekly allowance is essentially unused:
- Settings > Usage: Current session ~3%, Weekly (all models) 0%, Weekly Fable 0%.
- Claude Code /status → View usage: Session ~3%, Weekly (7 day) 0%, Weekly Fable 0%.
There is also a reset-time inconsistency: Settings > Usage shows the weekly
(all models) meter resetting today (Sat 8:59pm ET), a full week before the
block's stated reset (Jul 25, 9:00pm ET).
Critically, plain claude.ai (web browser) on the SAME account works normally.
The general subscription allowance clearly has capacity — only the Claude Code
surface is locked out, and no visible meter reflects the limit being enforced.
/status: Login method = Claude Max account (no API key), Model opus[1m]
(claude-opus-4-8[1m]). Same block also seen on claude-fable-5. Both
ANTHROPIC_API_KEY and CLAUDE_CONFIG_DIR are unset.
What Should Happen?
Either Claude Code should honor the actual weekly usage (0% used) and allow
requests, matching what claude.ai web allows on the same account; or, if a
Claude Code-scoped weekly limit is genuinely being enforced, it should appear
as its own meter in Settings > Usage and /status with a consistent reset time,
so the block is explainable instead of appearing at 0% reported usage.
Error Messages/Logs
You've hit your weekly limit · resets Jul 25, 9pm (America/New_York)
You've hit your weekly limit · resets in 6d
(Claude Desktop) Usage limit reached — resets Sat Jul 25, 9:00pmSteps to Reproduce
- Sign in to Claude Code with a Claude Max (20x) subscription (no API key set).
- Open Claude Code in the VS Code integrated terminal (WSL2 Ubuntu).
- Send any single trivial prompt (e.g., "hello") with no subagents.
- Observe the request is blocked: "You've hit your weekly limit · resets Jul 25".
- Run /status → open "View usage": Weekly (7 day) shows 0% used.
- Open Settings > Usage on claude.ai: Weekly (all models) 0%, resetting today
(not Jul 25).
- Open plain claude.ai in a browser on the same account: chat works normally.
Claude Model
Not sure / Multiple models
Is this a regression?
I don't know
Last Working Version
N/A
Claude Code Version
2.1.214
Platform
Other
Operating System
Ubuntu/Debian Linux
Terminal/Shell
VS Code integrated terminal
Additional Information
- The same account works on claude.ai web, proving the subscription allowance
is not exhausted; only the Claude Code surface is blocked.
- Claude Desktop "Code"/Cowork are also blocked, with the banner
"usage limit reached · shares limit with Claude Code".
- Session ID: 6a330222-3f72-448f-ab67-e647abda9660
- Screenshots attached: Settings > Usage (weekly 0%), /status output,
View usage panel (Weekly 7-day 0%), and the block banner.
3 Comments
Adding a related usage-ledger / entitlement-reconciliation signal, kept public-safe.
The direct symptom here is Claude Code enforcing a weekly-limit block on a Max 20x account while every visible usage meter shows roughly 0% weekly usage and claude.ai web still works on the same account. My unresolved case is a different Claude billing surface: manual prepaid/discounted usage-credit purchases fail or fail to commit correctly, while automatic usage-credit reload billing can still charge successfully from the saved billing setup.
The shared failure class is server-side billing and usage state diverging from customer-visible state. In this thread, visible usage meters, reset timing, Claude Code/Cowork scoped enforcement, subscription entitlement, account/session billing source, and support-visible state appear inconsistent. In my case, manual purchase state, auto-reload state, payment confirmation/commit, credit-ledger provisioning, and support-visible billing state diverge.
Actionable owner request: please route this class of issues to a human Billing Platform / usage-ledger / entitlement / Claude Code limit-enforcement owner who can inspect visible usage meters, hidden Claude Code-scoped limit buckets, reset-window state, account/session billing source, subscription entitlement, credit-ledger state, and support-visible billing state together before treating the block as true customer usage exhaustion.
Privacy-sanitized public incident hub:
https://coolak.github.io/anthropic-claude-billing-incident/
Reconciliation matrix:
https://coolak.github.io/anthropic-claude-billing-incident/reconciliation-matrix.html
I am intentionally not posting card details, bank names, payment object IDs, invoice IDs, private support identifiers, raw logs, screenshots, payment URLs, token values, exact private balances, or private support-thread text here.
Usage limit reached at 46% session — Claude.ai blocked while Claude Code works simultaneously
---
Description:
Account: Pro plan
Date/Time: August 18, 2026 — afternoon (London time)
Bug:
Claude.ai chat is showing a "usage limit reached" error, blocking all conversations. At the exact same moment, Claude Code CLI is working normally on the same Pro account.
Usage stats shown on the dashboard at time of error:
Expected behavior: At 46% session usage and 9% weekly usage, no limit should be reached.
Actual behavior: Claude.ai chat is blocked with a usage limit message. Claude Code CLI works fine simultaneously, which proves the account itself is not at capacity — this is an inconsistency on Anthropic's side.
Steps to reproduce: Unknown — I had barely used Claude today before the error appeared. I did not use Claude Code before the error; I only launched it after being blocked on claude.ai, which is how I discovered the inconsistency.
Request: Please investigate the session tracking logic — it appears the limit being blocked on claude.ai, which is how I discovered the inconsistency.
Request: Please investigate the session tracking logic — it appears the limit trigger is firing incorrectly, independently of actual usage.
Yesterday I wasn't able to use Claude Chat after only 9 messages sent( and the progression barre said I had used 53% of the session). Today I sent even less messages than that, and again, it says I have reached my limits and need to wait 4h.. I sent 5 messages in total (and not even complicated ones). On the session status, it says I have used 39% of my quota...... There is something wrong.