Anthropic Formal Complaint Public Version
Archive Header
Show metadata
- document_type
- essay
- title
- Anthropic Formal Complaint Public Version
- date
- 2026-06-12
- language
- en
- author
- Wang Xiao
- source_layer
- The Uncertain Future
- status
- public_archive
- canonical_route
- /uncertain-future/anthropic-formal-complaint-public-version
- source_url
- https://medium.com/@wangxiao8600/anthropic-formal-complaint-public-version-en-v1-8b3c7d24aa69
- intended_use
- This document should be read as a public author archive copy in The Uncertain Future, preserving Wang Xiao's time-specific structural judgment on AI, society, protocol, or structural change while retaining external publication links.
- not_for
- This document should not be treated as formal technical proof, legal advice, investment advice, career advice, external certification, or a complete statement of OathAI's current method layer.
- key_terms
- The Uncertain Future 路 Structure
- related_pages
- The Uncertain Future 路 Glossary
An unfair structure requires structural correction
Note
The text below is the public version of the complaint I have already formally sent to Anthropic.
This public version preserves the structure, order, and substantive wording of the original complaint. It has not been compressed or rewritten for summary purposes. The only changes are redactions of sensitive fields involving personal identifiers, account identifiers, invoice and receipt numbers, tax identifiers, and related identifying information.
The purpose of publishing this text is not to seek a refund, individualized compensation, or to substitute for the formal complaint process. On the contrary, publication takes place only after the internal complaint path has already been documented and a formal written notice has already been sent. The purpose of publication is to present, in full, my formal claims concerning Anthropic's handling of expired API paid credits, the relevant term structure, and billing-transparency deficiencies.
Account email: [REDACTED] Organization ID: [REDACTED] Invoice ID: [REDACTED] Receipt ID: [REDACTED] Purchase date: 2025-06-04 Billing country / VAT jurisdiction: Portugal / EU VAT Tax ID / VAT number shown on invoice: [REDACTED] Amount paid: USD 24.60 Paid credits: USD 20.00 VAT: USD 4.60 Visible usage after CSV reconciliation: USD 5.30 Remaining paid credits marked expired: USD 14.70
Hello,
As a preliminary matter, I state the following clearly: I do not intend to accept any refund, compensation, reissued credits, account adjustment, or any other form of individualized relief. Nor do I intend to accept any settlement arrangement that would substitute private handling for a public apology, amendment of the relevant terms, and repair of the billing mechanism. The sole purpose of this complaint is to require Anthropic to provide a formal response to the unfair term structure and billing-transparency deficiencies described below, and to take structural corrective measures. This statement does not mean that I refuse to receive a formal written response from Anthropic; what I refuse is any individualized relief or private handling offered as a substitute for structural correction of an unfair structure.
I am formally submitting a complaint regarding the Anthropic API Usage Credits that I purchased on June 4, 2025.
On June 4, 2025, I paid a total of USD 24.60, consisting of USD 20.00 in paid API credits and USD 4.60 in VAT. After manually aggregating and deduplicating the monthly Cost CSV records exported from the Anthropic Console for the period from June 2025 through June 2026, I confirmed that the visible actual usage of those credits was USD 5.30; based on the USD 20.00 paid-credits principal, the remaining paid credits that were marked as expired and then wiped out amounted to USD 14.70.
I also note that the original transaction included USD 4.60 in VAT. If prorated in proportion to the unused paid credits, the VAT corresponding to the unused portion would be approximately USD 3.38. This VAT-related issue forms a separate part of my complaint concerning transaction characterization and billing transparency.
My complaint is not limited to the fact that the remaining paid credits were wiped out. It also concerns the following issues: 1. the unilateral extinguishment of remaining paid credits purchased with real money; 2. the unfair structure of the expiration, top-up, and billing terms governing paid credits; 3. the lack of a transparent and user-verifiable billing record mechanism, because the current usage-history interface allows a custom query range of no more than 31 days, effectively forcing users to download CSV files month by month, merge them, deduplicate them, and reconstruct the lifecycle of a full year of credits in order to complete a proper verification.
I do not seek a cash refund. I reserve all rights to seek further remedies depending on Anthropic's response and the outcome of any regulatory review.
Because the issues described above, namely the unilateral wipeout of remaining paid credits after one year and the 31-day query limit that makes it difficult for users to verify the full credit lifecycle, do not arise from an isolated account-specific malfunction but from standard terms and a standardized user-facing interface applicable to relevant users, any individual refund, credit adjustment, or private settlement offer would be insufficient to address the underlying structural unfairness and transparency deficiencies reflected by this matter.
To resolve this complaint in a manner I consider adequate, Anthropic must:
I. Issue a formal public apology.
Anthropic must issue a formal public apology. That apology must be clear, substantive, and directly visible to affected users.
For such an apology to be meaningful and adequate, it must meet the following conditions: 1. it must explicitly acknowledge that remaining paid credits purchased with real money were directly wiped out after one year; 2. it must explicitly acknowledge that the current billing and usage-history interface, including the 31-day custom query limit, failed to provide sufficient transparency for users to verify the full lifecycle of their credits; 3. it must be published through Anthropic's principal official communication channels directed to API users and displayed in a manner sufficient to ensure that affected users actually receive notice of it; the relevant statement page must remain publicly accessible and must not be made less discoverable through technical means; 4. it must not use language that shifts responsibility onto users by suggesting "confusion," "misunderstanding," "inattention," or a "failure to monitor the account."
II. Amend the unfair expiration, top-up, and billing terms governing paid credits.
The relevant terms must no longer allow the platform, in circumstances where a user has purchased credits with real money and completed payment, to unilaterally wipe out the remaining paid balance solely on the basis of a pre-set one-year expiration period, without providing a fairer, more transparent, and more reasonable mechanism.
At a minimum, the following corrections must be made: 1. Anthropic must stop relying on the concept of "internal credits with no cash value" to weaken or negate the transaction's character as a real commercial transaction for which payment was collected and VAT was charged; 2. Anthropic must stop relying on standard-form contract terms to automatically clear or extinguish remaining paid balances after one year without providing sufficient, clear, and continuous advance notice and a workable path for users to address the issue before expiration; 3. Anthropic must provide a consistent, clear, and non-misleading explanation of the legal and transactional nature of remaining paid credits; 4. Anthropic must remove or amend any term that, in substance, results in the unilateral confiscation of balances that consumers have already paid for.
III. Repair the billing and account-history interface and provide a verifiable, complete credit ledger.
Anthropic must provide a transparent, user-accessible, and fully verifiable credit ledger that allows users to directly view, within one continuous and comprehensible record system: 1. the purchase date of each credit purchase; 2. the purchase amount of each credit purchase; 3. the actual usage during each relevant period; 4. the remaining balance; 5. the expiration date and expiration event; 6. the full process by which the balance changed over time.
To satisfy this requirement, Anthropic must not continue to maintain a billing structure under which users are forced to download CSV files month by month, merge them, and deduplicate them in order to verify how a one-year paid credit balance was used, how much remained, and how it expired. In particular, Anthropic must not continue to limit the custom query range to no more than 31 days while at the same time allowing one-year paid credits to run through a full lifecycle and ultimately be wiped out.
Please confirm receipt of this complaint and provide a case number. Please provide a substantive written response to the above matters within 14 calendar days. If Anthropic does not provide a substantive written response within that period, or expressly refuses to take corrective action regarding the matters described above, I will escalate this matter further by submitting formal complaints to the European Consumer Centres Network (ECC-Net) and to the competent Portuguese consumer-protection authorities, including the Dire莽茫o-Geral do Consumidor (DGC) or other appropriate authorities, together with the invoice, proof of payment, monthly Cost export records, reconciliation results, and relevant interface evidence.
Sincerely, Wang Xiao
https://medium.com/@wangxiao8600/anthropic-formal-complaint-public-version-en-v1-8b3c7d24aa69
https://open.substack.com/pub/wangxiao8600/p/anthropic-formal-complaint-public?r=5lo0u8&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true
About the Author
Wang Xiao is an AI protocol architect, author of System and Freedom, creator of Danbing AI Protocol / SLAPS Framework, and initiator of OathAI.
His work focuses on human-AI co-creation, protocol governance, semantic anchoring, and long-term knowledge continuity, exploring how human knowledge and collaborative structures can be preserved, calibrated, and inherited in the AI era.
Disclaimer
This essay reflects the author's current observations and methodological reflections based on personal practice, research, and human-AI collaboration experience. The related Danbing / SLAPS / OathAI methods are still being organized and evolved. Their practical effects may vary depending on the user's background, task context, model capability, execution environment, and level of commitment.
This essay does not constitute legal, investment, medical, career, or technical implementation advice or guarantee. Readers who apply these methods in real projects should make independent judgments based on their own circumstances and take responsibility for specific outcomes.