[BUG] ANTHROBUG/systemic billing defect: $600 charged, account on Free, credits never applied — Fin AI admits no human escalation available

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?

NOT A DUPLICATE OF #71821 — see full reasoning at the bottom of this report. This is a distinct systemic billing defect with written admission from Anthropic's own AI support agent.

I have been charged $600 across three payments for Anthropic services, but my account shows the Free plan, a $0.00 credit balance, and no usage history. The service I paid for was never delivered.

Payment details:

• Aug 21, 2026 — $200 — Max 20x subscription — Status: Paid
• Sep 10, 2026 — $200 — Usage credits — Status: Paid
• Sep 19, 2026 — $200 — Usage credits — Status: Paid

Total charged: $600

Current account status:

• Plan: Free
• Credit balance: $0.00
• Usage history: Completely empty — no record of any credit consumption
• Workspace created: September 22, 2026 — AFTER the Sep 10 and Sep 19 charges

This is not an isolated case. The following existing issues describe the identical pattern:

• #57594 — Fin AI checks API organization instead of consumer subscription
• #91896 — Paid Pro reverted to Free, Fin confirmed known issue, no resolution after 2 weeks
• #86505 — Paid Max 20x ($209.10) with no subscription created
• #35340 — Paid Max plan not reflected, account on Free
• #56895 — Max payment charged but account reverted to Free
• #67407 — Account downgraded due to billing error
• #68257 — Max subscription silently downgraded 2 days after payment
• #45890 — claude.ai subscription not applied to account

Anthropic's own status page (status.claude.com) has documented related incidents:

• January 9, 2026 — "New Subscription Activation & Payment Issues" — "some new customer subscriptions are charging customers without properly granting subscription access, which is also leading to accidental double or triple payments for some customers."
• January 29, 2026 — "Delays in purchasing additional API credits"
• September 2, 2026 — "Delays in credit purchases" — "users who reached a balance of zero usage credits saw delays in the availability of newly purchased credits" (affected credits purchased from Sep 5, 2026 onward)

---

NOT A DUPLICATE OF #71821

This is not a duplicate of #71821. While the surface symptom is similar (paid plan showing as Free), there are key differences:

1. Different plan and scope. #71821 involved a Max 5x plan (single charge of $91.28). My case involves a Max 20x plan PLUS two $200 credit purchases — three charges totaling $600.

1. Written admission from Anthropic's own AI support agent. In my case, Fin AI Agent has confirmed in writing that:

• The September charges occurred before my Workspace was created on September 22.
• This is "inconsistent with how usage credit purchases should function."
• The case "requires investigation beyond what the standard troubleshooting can resolve."
• "Escalation is not available through this channel." Fin AI cannot transfer conversations, notify internal teams, or trigger escalation workflows.

1. #71821 was closed as "invalid" without resolution. Mine is not about Claude Code behavior — it is a systemic billing/entitlement defect that affects multiple users (see #57594, #91896, #86505, #35340, #56895, #67407, #68257, #45890).

1. This is a documented, recurring pattern. It is now the subject of a class-action lawsuit and multiple FTC complaints. Anthropic's BBB rating is F with 44 unanswered complaints. Trustpilot rating is 1.5 / 5.

I am filing here because this repository is the only public channel with active human monitoring for billing defects of this nature.

Please do not close this as a duplicate. It is a distinct case with distinct evidence.

---

Full evidence package (invoices, receipts, screenshots, Fin AI correspondence, complete ANTHROBUG disclosure) is available on request to Anthropic staff.

Account identifiers (email, transaction IDs, Organization ID) are available in the evidence package for Anthropic verification, but are not published publicly to avoid scraping and phishing.

What Should Happen?

Either:

1. Activate the Max subscription I paid for on my consumer account (claude.ai), OR
2. Refund $600 in full — the $200 Max subscription charge plus the $400 in credits that were never delivered or applied.

Additionally:

1. Human review of this case by a billing specialist — Anthropic's own Fin AI Agent has confirmed in writing that this case requires investigation beyond standard troubleshooting, and that no escalation path currently exists.
2. Written acknowledgment of the unauthorized charges if neither activation nor refund is possible.
3. A documented path to human escalation for consumer accounts — because as Fin AI itself stated in writing: "escalation is not available through this channel... I do not have the ability to transfer conversations, notify internal teams, or trigger escalation workflows."

Error Messages/Logs

No application error. The defect is silent — the account shows "Free plan" with an "Upgrade Plan" button, while three successful payments are visible in billing history.

However, Anthropic's own AI support agent (Fin AI Agent, support@mail.anthropic.com) issued the following statements **in writing** after five escalation requests:

---

"I cannot provide a ticket number, as our email-based support system does not generate ticket numbers in the traditional sense. This email thread itself serves as your support record."

"I cannot confirm escalation to a human billing specialist because **escalation is not available through this channel**. As an AI support agent, I provide information and guidance based on our help documentation, but **I do not have the ability to transfer conversations, notify internal teams, or trigger escalation workflows**."

"There is a **clear discrepancy** between your payment history and your current account status that **requires investigation beyond what our standard troubleshooting can resolve**."

"The September 10 and September 19 charges occurred **before your Workspace was created on September 22**, which is **inconsistent with how usage credit purchases should function**."

"Your documentation is thorough and the issue you've described **falls outside normal subscription behavior patterns** documented in our help resources."

---

**Summary:** Anthropic's own support agent has confirmed in writing that:
- A clear discrepancy exists between payments and account status.
- The case requires human investigation.
- The charges are inconsistent with documented system behavior.
- **No escalation path to a human exists.**

Steps to Reproduce

This is a billing/account state defect, not a code bug. The following steps reproduce the account state:

1. Purchase a Claude Max 20x subscription on Aug 21, 2026 — payment of $200 processes successfully, invoice status shows "Paid."
2. Purchase $200 in usage credits on Sep 10, 2026 — payment processes successfully, invoice status shows "Paid."
3. Purchase $200 in usage credits on Sep 19, 2026 — payment processes successfully, invoice status shows "Paid."
4. Create a Claude Console Workspace on Sep 22, 2026.
5. Sign in to claude.ai and navigate to Settings → Billing.

Expected behavior:

• Max 20x subscription active
• $400 in usage credits available on the account
• Usage history reflecting the credit purchases
• No "Upgrade Plan" button (since a paid plan is already active)

Actual behavior:

• Plan shows Free
• Credit balance shows $0.00
• Usage history is completely empty — no record of any credit consumption
• "Upgrade Plan" button is displayed, as if no paid subscription exists
• No failed payment notifications, no subscription status message, no error of any kind

Note: The Sep 10 and Sep 19 charges occurred before the Workspace was created on Sep 22 — which is inconsistent with how usage credit purchases should function, as confirmed in writing by Anthropic's own Fin AI Agent.

Claude Model

None

Is this a regression?

No, this never worked

Last Working Version

N/A

Claude Code Version

N/A — this is a claude.ai consumer billing issue, not a Claude Code bug. Filed here because this repository is the only public channel with active human monitoring for billing defects (see #57594, #91896, #86505, #35340).

Platform

Other

Operating System

Windows

Terminal/Shell

Other

Additional Information

Important context:

• This is not a Claude Code regression. It is a server-side billing/account state defect affecting claude.ai consumer subscriptions.
• Filed in this repository because it is the only public channel with active human monitoring for billing defects of this nature — as evidenced by existing issues #57594, #91896, #86505, #35340, #56895, #67407, #68257, #45890.
• The pattern is systemic: multiple users, multiple months, identical symptoms.
• Anthropic's own status page has documented related incidents (Jan 9, Jan 29, Sep 2, 2026).

Anthropic's own AI support agent (Fin AI) has confirmed in writing that:

• A clear discrepancy exists between payments and account status.
• The case requires human investigation.
• No escalation path to a human is available.
• Fin AI cannot transfer conversations, notify internal teams, or trigger escalation workflows.

Full evidence package (invoices, receipts, screenshots, Fin AI correspondence, complete ANTHROBUG disclosure) is available on request to Anthropic staff.

Account identifiers (email, transaction IDs, Organization ID) are available in the evidence package for verification, but are not published publicly to avoid scraping and phishing.

This issue is par t of a documented, recurring pattern that is now the subject of:

• A pending class-action lawsuit against Anthropic regarding misleading Max plan advertising
• Multiple FTC complaints
• A BBB rating of F with 44 unanswered complaints
• A Trustpilot rating of 1.5 / 5

Request: Please do not close this as a duplicate of #71821. It is a distinct case with distinct evidence. Please escalate to a human for review.