Terms of Service
This is draft v0.1. The official version will be published after legal review; it does not constitute contractual terms before taking effect.
01Scope
These Terms govern your access to and use of the AI inference cloud services provided by Plenum Cloud at plenumcloud.ai, including the API, console, dedicated endpoints, private deployment and compute onboarding (collectively, the "Services"). Plenum Cloud is an OPTIWAVE platform for large-model and compute services operated by Guangyuan Weilai (Shanghai) Technology Co., Ltd. ("we", "us"). plenumcloud.ai serves customers outside mainland China; customers in mainland China should use plenumcloud.com. Accounts and credit are not shared between the two sites.
By registering an account, obtaining an API key or otherwise using the Services, you agree to these Terms. If you use the Services on behalf of an organisation, you represent that you are authorised to bind it.
02Accounts and API keys
New accounts are activated after our review. You are responsible for safeguarding credentials and API keys; a key's plaintext is shown only once at creation and should be revoked immediately in the console if leaked. All usage and charges made with your credentials are your responsibility.
03The Services
Services are offered in tiers: Serverless / Batch (pay-as-you-go), Elastic Endpoints, Dedicated Endpoints (exclusive capacity with service-level commitments) and Private / Appliance delivery. Specifications, quotas, rate limits and service levels for each tier are defined in the order, contract or product description in the console.
For requests routed to third-party model providers (for example OpenAI), model capability, versioning, availability and content policies are determined by that provider; we are responsible for routing, metering and failover, not for the content of third-party outputs. When using such models you must also comply with the provider's published usage policies.
04Billing and payment
Prices are in US dollars. Usage-based services are billed according to platform metering, which counts input, cached-input, output and reasoning tokens separately at their respective prices. Dedicated endpoints are billed by capacity × time regardless of actual usage. Invoices and usage records are available in the console.
Top-ups are prepaid credit (US$50–US$50,000 per order) paid by bank wire to the account we publish, and are credited once finance confirms receipt; bank and intermediary fees are borne by the sender. Prepaid credit may only be used to offset fees for the Services; it bears no interest and cannot be transferred, withdrawn or moved between accounts. Signup credit is non-refundable. Refunds, where requested in writing, are limited to the unconsumed top-up portion and are returned via the original channel after verification, net of bank charges actually incurred. Prices exclude applicable taxes, which are your responsibility.
Prices may change with prior notice via the console or email; contracted dedicated and private deployments follow the contract price. We may suspend the Services for overdue accounts until payment is settled.
05Compute onboarding and settlement
Compute providers whose endpoints pass review earn according to actual service: shared traffic at the per-token payout rate agreed at approval, dedicated instances at the agreed hourly rate, aggregated by calendar month. Earnings become withdrawable 7 days after month end (minimum US$500 per request), paid by wire with fees deducted from the payment. Providers are responsible for their own tax obligations and must supply an invoice or equivalent document on request. The payee account holder must match the registered entity. Amounts in dispute (fabricated nodes, policy breaches, customer complaints) may be frozen until resolved.
06Acceptable use
You must comply with the laws applicable to you and to the provision of the Services, and with the usage policies of the third-party model providers you use. You may not use the Services to generate or distribute unlawful, harmful, fraudulent or infringing content, infringe intellectual property, privacy or trade-secret rights, engage in activities that endanger network security, or use the Services for purposes or in jurisdictions prohibited by applicable export-control or sanctions laws.
You may not circumvent rate limits, quotas or safety mechanisms, nor conduct unauthorised load testing, reverse engineering or resale of the Services. We may throttle, refuse, suspend or terminate the Services for violations.
07Data
Content you submit and the outputs generated belong to you. We never use customer data for model training and enforce technical isolation. The global site's control plane, accounts and usage ledger are hosted in the United States; request content is forwarded only to the compute node or third-party model provider serving the request, subject to that provider's terms. The usage ledger records only token counts and metadata, not request or response bodies. See the Privacy Policy.
08Availability and service levels
Availability commitments and remedies for dedicated endpoints and private deployments are set out in the contract. Serverless / Batch services are provided on a best-effort basis without availability commitments. Parts of the Services run on third-party and idle compute and may experience node interruptions; the platform mitigates impact through health probing and failover but does not guarantee zero interruption.
09Disclaimer and limitation of liability
The Services are provided "as is" without warranties of any kind, express or implied, to the extent permitted by law. To the maximum extent permitted by law, we are not liable for indirect losses, lost profits or data loss arising from use of or inability to use the Services; our total liability is limited to the fees you actually paid for the relevant Services in the twelve months preceding the dispute.
10Termination
You may close your account at any time. We may suspend or terminate the Services for material breach, prolonged non-payment or legal requirements, with reasonable notice except in emergencies. Fees already incurred remain payable after termination, and your data will be handled per the Privacy Policy.
11Changes
We may update these Terms; updates are dated on this page and announced via the console or email, with at least fifteen days' notice for material changes. Continued use after the effective date constitutes acceptance. The English version of these Terms prevails; the Chinese text is a courtesy translation.
12Governing law and disputes
These Terms are governed by the laws of the Hong Kong Special Administrative Region, excluding its conflict-of-laws rules. Disputes shall first be resolved through good-faith negotiation; failing that, they shall be referred to and finally resolved by arbitration administered by the Hong Kong International Arbitration Centre (HKIAC) under its rules in force at the time, seated in Hong Kong and conducted in English.
13Contact
For questions about these Terms, contact us through the channels published in the website footer (Guangyuan Weilai (Shanghai) Technology Co., Ltd., Room 1709, Canxing Building, 158 Longqi Road, Xuhui District, Shanghai, China).