Terms

Terms of use

These terms apply when you visit the Flonno website or use an API service operated by its publisher.

1. The service

Flonno provides an account dashboard for creating scoped API tokens and an OpenAI-compatible interface to the configured Qwen3.8 Flash 27B inference route. Available models, request limits, and supported features may change.

2. Accounts and API keys

Keep API keys private and use them only in server-side applications. A key can access only the endpoints and models granted when it was created, subject to expiry or revocation. You are responsible for activity made with keys issued to your account and should revoke any key that may have been exposed.

3. Model requests and providers

Flonno runs model inference itself on its own GPU inference system in Switzerland. Requests pass through our gateway to our RTX 3090 runtime, without routing to another AI API provider. No automatic external-provider fallback is configured. Hosting infrastructure remains involved; other service data is not guaranteed to stay in Switzerland. Qwen3.8 Flash 27B is an optimized, additionally quantized derivative of Qwen3.8-27B; quality may differ from the unquantized model. The API supports text, inline image understanding and tools, with a shared 131,072-token input/output context and at most 32,768 output tokens. It does not provide TEE isolation or attestation. Review the privacy notice and obtain any necessary processor agreements and transfer information before submitting personal data.

4. Acceptable use

You must not use Flonno to violate the law, infringe another person’s rights, compromise systems, evade provider restrictions, or create or distribute harmful content where prohibited. Follow the requirements of every model and service you connect.

5. AI outputs and availability

AI outputs can be inaccurate, incomplete, or unsuitable for your purpose. Review outputs before relying on them, especially for consequential decisions. The service is provided without a promise of uninterrupted availability or error-free operation.

6. Prepaid usage and payments

Sign up and verify your primary email to receive $1 in promotional API credit automatically on opening your dashboard. No credit card or subscription is required. This is one evaluation allowance per person, including any previous pilot credit; duplicate accounts, repeated claims and abuse are excluded. Promotional credit has no cash value, is not transferable or refundable, and does not expire. Once exhausted, further billable requests require an optional paid top-up; there is no automatic charge. We retain a keyed hash of the verified email to prevent repeat claims, including after account deletion.

Prepaid credit is purchased in USD through Stripe Checkout and applied after payment confirmation. The gateway reserves credit before inference and settles against reported token usage. Requests with missing usage or interrupted responses may remain pending for reconciliation. The published Qwen rate for a new account is $0.05 per million input tokens and $0.80 per million output tokens, including reported thinking tokens. An account created before 10 November 2026, 08:01 UTC keeps $0.03 per million input tokens and $0.25 per million output tokens, including when its wallet is created after that time, for as long as that stored rate remains in effect. An account created at or after that time is billed at the published rate. Cached input is charged at the input rate stored for that account. There is no separate cached-input price. A reservation is not a second charge: unused reserved credit is returned on settlement. Ambiguous interrupted requests and missing usage require reconciliation; no automatic release deadline is promised. Review displayed prices before use; prices may change and no minimum advance-notice period is currently promised. Each admitted request retains its quoted input/output rates for settlement. API use and top-ups may be paused while a payment dispute is open or a lost dispute requires reconciliation. Contact the publisher about billing errors, disputed payments, unused credit or refund requests. These terms do not exclude mandatory cancellation, refund or consumer rights.

For a billing review, unused-credit refund request or capacity review, contact [email protected]. Include the payment or reservation reference and request time, never an API key or prompt content. A refund request is reviewed individually; no general automatic refund promise is made. Larger limits and any service commitments require a separate agreement. Current capacity and recovery limits are described on the status page.

7. Changes

These terms may be updated as the service changes. The version published here applies from its stated update date. Mandatory consumer protections and other non-waivable rights under applicable law remain unaffected.

Last updated 10 November 2026.