| Operator / brand | metaproxy |
|---|---|
| Covered products | https://metaproxy.agency, the console, API gateway (https://api.metaproxy.agency/v1), documentation, and related services (collectively, the “Service”) |
| Version | v1.0 |
| Effective date | July 31, 2026 |
| Support | support@metaproxy.agency |
| Legal | legal@metaproxy.agency |
Important notices
- metaproxy is an independent third-party model aggregation and routing gateway. It has no agency, authorization, or official partnership with OpenAI, Anthropic, Google, xAI, or other upstream providers. Their trademarks belong to their owners and are referenced only to identify available models.
- The Service is provided “as is” and “as available.” Model output may be inaccurate and must not be used for medical, legal, financial, or other high-risk decisions without qualified human review.
- These terms include limitations of liability, disclaimers, and an acceptable-use policy that may materially affect your rights.
- Online payments are handled by third-party processors, currently including Creem. Credits are used to call the Service under the billing rules below.
Parties and scope
These Terms of Service are a binding agreement between metaproxy (“we”, “us”) and the person or entity that visits, registers for, or uses the Service (“you”).
The Privacy Policy, published pricing and billing explanations, and incorporated upstream acceptable-use policies form part of this agreement.
If you use the Service for an organization, you represent that you are authorized to bind it to these terms.
Service description
metaproxy provides:
- A unified OpenAI-compatible API gateway and base URL;
- API-key management, a model catalog, and usage metering;
- A console for usage, logs, bills, balances, and settings;
- Integration documentation, example scripts, and public landing-page information;
- Prepaid credit top-ups and usage records.
We do not train or own the underlying large language models. We act as a technical intermediary for routing, forwarding, metering, billing, and observability. Upstream providers determine model capability, availability, and output.
When you call an upstream model, you must also follow that provider’s current terms and acceptable-use policy. For that model, the stricter rule controls in case of conflict.
Eligibility
- You must be at least 18 years old and legally capable of entering this agreement. The Service is not offered to minors.
- You must not be in a jurisdiction subject to applicable comprehensive sanctions or export controls, provide the Service to a sanctioned party, or use it for a prohibited end use.
- You are responsible for complying with laws applicable to data, privacy, AI, telecommunications, and cross-border payment where you operate.
- You must not use a false identity, misstate your purpose, or evade eligibility or compliance controls.
Accounts and security
- Provide accurate, complete registration and contact information and update it when it changes.
- Protect account credentials, passwords, API keys, and OAuth credentials. To the extent allowed by law, activity through your account or key is treated as yours unless caused by our systemic error.
- Do not expose complete API keys in public repositories, screenshots, frontend bundles, or communities. You are responsible for credit consumed, abuse, or loss caused by a leaked key.
- Without written permission, do not sell, rent, lend, or transfer accounts or keys, or repackage the Service as a resale gateway, account pool, or competing relay. Authorized channels are excepted.
- For security, risk, overdue payment, abuse, upstream requirements, or law, we may warn, rate-limit, suspend, disable, or terminate an account, key, or feature, including without advance notice where necessary.
Acceptable use and prohibited conduct
Use the Service only for lawful, proper purposes permitted by these terms and upstream policies. You must not use or allow others to use it to:
- Violate applicable law, regulation, or an upstream acceptable-use policy;
- Harm critical infrastructure; create or distribute malware; conduct unauthorized intrusion, DDoS, or large-scale attacks;
- Create or distribute child sexual abuse material, including AI-generated material, or any sexualized content involving minors;
- Support terrorism, violent extremism, hate, harassment, large-scale disinformation, or election manipulation;
- Infringe intellectual property, privacy, publicity, or other lawful rights;
- Use output directly for regulated high-risk decisions such as medical diagnosis, legal advice, credit approval, or employment without qualified human review and required disclosures;
- Evade rate, geographic, or security controls, or conduct unauthorized jailbreak, distillation, or adversarial attacks against the platform;
- Use the Service to train a competing model at scale or collect inputs and outputs for that purpose;
- Commit fraud, abuse chargebacks, create artificial transactions, manipulate referral rewards, or harm the platform or other users.
We may, but are not obligated to, use automated abuse and compliance screening and may block, limit, or reject high-risk requests. Serious violations may result in permanent account termination without refund of unused credit, except where mandatory law requires otherwise.
If you believe an enforcement decision is wrong, request human review within 14 days of learning of it at support@metaproxy.agency.
User content, output, and training
- “Input” includes prompts, messages, files, and parameters you submit; “Output” is content returned by a model. To the extent allowed by law, you retain rights you lawfully hold in Input and Output. We do not claim ownership, subject to upstream terms.
- You grant us a limited, non-exclusive, revocable license only to receive, route, process transiently, meter, bill, troubleshoot, and apply risk controls as needed to provide the Service.
- We do not use Input or Output to train, fine-tune, or evaluate any AI model and do not sell conversation content.
- We do not retain complete Input or Output bodies long-term by default. We retain usage metadata needed for billing and security as described in the Privacy Policy.
- AI output can be inaccurate, outdated, biased, or contain unsafe code. You must review and test it and must not use it as the sole basis for high-risk decisions.
- You represent that you have the rights needed for Input and that Input and downstream use comply with law and third-party rights.
Billing, top-ups, and refunds
Prepaid credits
The Service uses prepaid credit. Add credit under the console and pricing-page rules before making API calls. Charges can depend on model, input, output, cached tokens, streaming status, and other published billing dimensions. Multiple keys can share one balance.
The “$” symbol denotes a dollar-denominated prepaid account unit, not USD cash, a foreign-currency deposit, or a withdrawable asset. Each ¥1 paid adds $1 of account balance. Credit cannot be transferred, sold, or used outside the Service.
Pricing
Model rates and savings claims relative to official prices are governed by the pricing page. Both model charges and official USD list prices are converted into account-balance units at the fixed reference rate of ¥7.3 per USD, so the recharge-ratio change does not alter the savings percentage. We claim savings only where a citable public USD provider price exists and our price is genuinely lower. We may change prices or model availability because of upstream cost, exchange rates, compliance, or supply. We will try to give advance console or email notice of material adverse changes, except for urgent security or upstream changes.
Payments
Third-party processors, currently including Creem, handle online payment. Their terms and privacy policies also apply during payment. We do not store complete sensitive card data.
Refunds
| Situation | Handling |
|---|---|
| General rule | Top-ups are generally final. Consumed credit is non-refundable. |
| Unused window | If there is no successful API call within 24 hours of a top-up and no bonus or promotion was received, you may request a refund of the unused amount actually paid. Processor fees and reasonable handling cost may be deducted. |
| Our error | Verified overcharges caused by an obvious pricing error or service failure will be refunded or restored as credit. |
| No refund | Consumed credit; promotional credit; balance remaining after termination for your breach; amounts subject to a unilateral chargeback before attempting resolution with us; or refunds prohibited by law or the payment channel. |
Email refund requests with the order number and supporting evidence to support@metaproxy.agency. We will investigate and respond within a reasonable number of business days. Raise transaction disputes within 30 days of the transaction.
Availability and changes
We use reasonable efforts to keep the Service stable but do not guarantee continuous availability, error-free operation, a particular latency, or long-term availability of a model, except under a separate written SLA. Maintenance, network failures, upstream throttling or withdrawal, payment-channel problems, attacks, compliance requirements, or force majeure may interrupt or degrade the Service, for which we are not liable to the extent allowed by law.
We may add, change, or remove models and features. We will try to provide a transition period for non-experimental model removals; experimental and preview models can change at any time.
Intellectual property
The site, console, brand, documentation, example code except portions expressly provided for integration, interface design, and platform-derived statistics are owned by metaproxy or their respective rights holders. Without permission, do not copy, mirror, reverse engineer, or use our brand in a confusing way.
OpenAI, Claude, Anthropic, Gemini, Grok, and other marks belong to their owners. Do not claim official authorization or agency from an upstream provider.
Privacy
We process personal information and usage data under the Privacy Policy. Your responsibilities as a controller or processor of third-party data do not transfer to us because you use the Service.
Disclaimers and limitation of liability
TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, THE SERVICE IS PROVIDED “AS IS” AND “AS AVAILABLE,” WITHOUT EXPRESS OR IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, NON-INFRINGEMENT, ACCURACY, OR CONTINUITY.
Except where mandatory law provides otherwise, we are not liable for loss caused by:
- Your mismanagement of a key or account;
- Your code, network, local tools, or third-party services;
- Changes, suspension, or termination by upstream models or payment channels;
- Force majeure, regulatory requirements, or your violation of these terms;
- Inaccurate, biased, or unsafe model output.
TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, OUR TOTAL AGGREGATE LIABILITY TO YOU WILL NOT EXCEED THE GREATER OF THE SERVICE FEES YOU ACTUALLY PAID AND CONSUMED DURING THE 12 MONTHS BEFORE THE EVENT GIVING RISE TO LIABILITY OR US$100. WE ARE NOT LIABLE FOR INDIRECT, INCIDENTAL, SPECIAL, PUNITIVE, OR CONSEQUENTIAL LOSS, INCLUDING LOST PROFIT, GOODWILL, OR DATA. Where local law does not allow part of this exclusion, the limitation applies only to the maximum extent it permits.
Indemnification
To the extent allowed by law, you agree to defend and indemnify us, including reasonable legal fees, against third-party claims, fines, or losses arising from your use of the Service, violation of these terms or upstream policies, infringement of third-party rights, or disputes with your end users.
Suspension and termination
You may stop using the Service and request account closure at any time. We may limit, suspend, or terminate the Service for breach, overdue payment, abuse, law, or upstream requirements. Termination does not affect payment obligations, indemnities, liability limits, or dispute provisions accrued before termination.
Changes to these terms
We may revise these terms for business, legal, or upstream changes. We will provide reasonable notice of material changes through the site, console, or email. Continued use after notice means acceptance. If you disagree, stop using the Service and request closure before the change takes effect.
Governing law and disputes
Disputes arising from these terms or the Service should first be discussed in good faith through legal@metaproxy.agency. If unresolved, either party may bring proceedings in a court with jurisdiction at the Service provider’s principal place of operation, except where mandatory law requires another forum.
If any provision is invalid, the remaining provisions continue in effect.
Contact us
| Support | support@metaproxy.agency |
|---|---|
| Privacy | privacy@metaproxy.agency |
| Legal | legal@metaproxy.agency |
| Website | https://metaproxy.agency |