1. Upstream providers and model services Use of GPT or image-related models is subject to OpenAI's terms, usage policies, safety policies, and regional restrictions. Use of other models or channels is subject to the relevant provider, distribution-channel, and legal requirements. Model names, sources, prices, quotas, and availability are shown in the dashboard or on the purchase page. Availability, output quality, context length, rates, pricing, data-processing rules, content review, abuse detection, regional restrictions, and continuity may be affected by upstream providers.
Service-Specific Terms
Effective: June 20, 2026. Last updated: June 20, 2026. When using StringRoute, users must also follow applicable upstream model-provider policies. The refund rules on this page apply to Webchat memberships, API balance top-ups, quota packages, enterprise plans, promotions, and related orders.
2. General refund principles AI services incur immediate compute usage, upstream model costs, payment-channel fees, exchange-rate costs, and risk-control costs. Unless this policy states otherwise, the purchase page provides otherwise, or applicable law requires otherwise, services that have been used, consumed, expired, incurred upstream costs, or been frozen for a violation are generally non-refundable. Resources are normally enabled immediately after a top-up, subscription, or quota-package purchase. Users should confirm plan contents, applicable models, validity, limits, price, and refund rules before paying.
3. Webchat memberships Webchat memberships cover personal chat, file analysis, image understanding, conversation history, PWA, and related features. A refund may be requested within 24 hours of a first purchase if advanced models, file analysis, image understanding, and other high-cost features have not been materially used. A full refund is generally unavailable after the subscription period has begun and material use has occurred. Subscriptions renew automatically and may be cancelled before the renewal date. A successful renewal is non-refundable unless applicable law requires otherwise. Future renewals may be cancelled at any time. Benefits remain available until the end of the paid subscription period.
4. API balance top-ups API balance is used for metered model calls. Amounts corresponding to consumed balance, completed calls, upstream costs, payment fees, channel fees, exchange losses, or risk-control costs are non-refundable. API balance top-ups are non-refundable. Rules for unused balances, minimum refund amounts, fees, refund periods, and refund methods do not apply unless applicable law requires otherwise or we separately agree in writing. If a user violates the Terms of Service, Fair Use Policy, anti-abuse rules, upstream policies, or applicable law, including through API abuse, API Key sharing or resale, multi-account arbitrage, fraud, chargebacks, attacks, or prohibited content, we may deny refunds, freeze balances, and deduct incurred costs and losses.
5. Quota packages, gifted quotas, and promotions Quota packages, discounted bundles, limited-time promotions, gifted quotas, free quotas, test quotas, and campaign benefits generally cannot be redeemed for cash, transferred, extended, or refunded unless the purchase page expressly states otherwise. Validity is shown on the purchase page, in the dashboard, or in the order details. Unused quota expires automatically without cash redemption, refund, or extension. Multiple quota packages may be combined and are consumed in order of expiry.
6. Enterprise plans Refund rules for enterprise plans, custom quotes, private deployments, dedicated channels, bulk quotas, SLAs, technical support, dedicated model routing, or custom integrations are governed by the signed order, contract, or addendum. If there is no separate signed agreement, this policy and the purchase-page terms apply.
7. Non-refundable situations Tokens have been consumed, a model has been called, output has been generated, or a high-cost feature has been used. User error, changed business needs, or failure to review pricing, plans, quotas, models, or regional restrictions. Inability to use the service caused by the user's network, device, browser, proxy, third-party system, or code. Service restrictions caused by prohibited content, API abuse, account sharing, quota resale, abusive refund conduct, chargebacks, or fraud. Free, gifted, test, or promotional quotas and benefits. The refund request period has expired. Payment-channel, exchange-rate, bank, card-issuer, tax, or third-party fees, unless law requires otherwise.
8. Service incidents and refund requests If a StringRoute issue makes the core service continuously unavailable for more than 48 hours, we may extend a membership, issue equivalent quota, or provide compensation based on actual impact. Standard Webchat, standard API balances, and non-enterprise plans do not include an SLA. Enterprise SLAs require a separate written agreement. Model slowdown, queues, rate limits, temporary upstream unavailability, model replacement, feature changes, price changes, regional restrictions, or security-policy rejection do not automatically create a right to a refund. To request a refund, email [email protected] with the registered email, order number, payment time and amount, reason, last four characters of the API Key, and supporting material. We will review the request within 10 business days. Settlement timing depends on the payment channel, bank, or card issuer. Approval is not guaranteed.