Privacy Policy

1. Summary

The rest of this policy sets out the detail. In short:

• We record metadata about each request — timestamp, model, token counts, latency, endpoint, identifiers, IP address and the amount charged — because we cannot bill, rate-limit, route or troubleshoot without it.
• We do not use your prompts or outputs to train models, and we do not sell personal data.
• To answer your request, we must transmit its content to the model provider serving the route you selected or that our routing selected. That provider’s own data practices then apply, and they differ between routes — see Section 6.
• If you sign in with Google, we receive only your Google account identifier and email address, and our use of that data follows Google’s Limited Use requirements — see Section 4.
• You can access, export, correct and delete your account data, and you can close your account at any time.

2. Scope and Controller

We are the controller of the personal data described in this policy. It applies to visitors to onefast.ai, to registered account holders, and to administrators and authorised users of organisation and reseller accounts.

Where you submit personal data of third parties inside your API requests, you are the controller of that data and we act as your processor, handling it only as described in Section 4 and only to deliver the Service.

This policy does not cover the practices of model providers, payment processors, identity providers or other third parties, each of which publishes its own policy.

3. Personal Data We Collect

Data you provide

• Account data: username, display name, email address, and a hashed password. We store passwords only as salted hashes and cannot recover them.
• Federated sign-in identifiers: if you sign in with Google, GitHub, LinuxDO or another supported identity provider, we receive and store the stable account identifier issued by that provider, and the email address and display name it releases to us. We do not receive your password for that provider.
• Security settings: if you enable two-factor authentication, the secret used to verify your codes.
• Billing data: your top-up history, credit balance, consumption, invoice-related details you supply, and the customer reference issued to us by our payment processor. We do not receive or store full payment card numbers.
• Communications: the content of support requests, bug reports and other correspondence you send us, including any attachments or log excerpts you choose to include.
• Organisation and reseller data: for accounts within a reseller hierarchy, the parent-account relationship, assigned pricing group, and any contact details a reseller records for its own customers.

Data collected automatically

• Request metadata, recorded once per billable request: timestamp, account and API key identifier, key name, model name, pricing group, input/output/cached token counts, amount charged, response time, whether the response was streamed, the API endpoint path, our internal request identifier, the upstream response identifier, and the IP address the request originated from.
• Error diagnostics: when a request fails, we record the error message returned to us, which is generated by the upstream provider or by our gateway. These messages describe the failure (for example rate limits, authentication failures, or content-policy rejections); in uncommon cases an upstream error message may quote a fragment of the offending request.
• Web session data: authentication cookies, and standard server-side records of requests to the web interface such as IP address, user agent, referring page and timestamp.
• Account region: when you register or sign in, we derive a country code (for example "DE") from the IP address of that request, or from the region in your browser language setting, and store only that two-letter code on your account; a country you choose yourself always takes precedence. We do not store the IP address for this purpose, nor any city-level or more precise location. It is used to support you and to understand which markets we serve.

Data we deliberately do not collect

We do not collect special categories of personal data (such as health, biometric, racial or ethnic, political or religious data) about you as an account holder, and we ask that you do not send such data to our support channels. We do not build advertising or behavioural profiles, and we do not purchase personal data from data brokers.

4. Google User Data

If you choose to sign in with Google, this section describes exactly what we receive and what we do with it. It applies in addition to the rest of this policy.

What we request and receive

We request only the “openid”, “email” and “profile” scopes. From the resulting profile we store your Google account identifier (the “sub” claim), your email address and its verified status. We use these solely to create your account, to sign you in on subsequent visits, and to contact you about the Service.

We never receive your Google password. We do not request access to Gmail, Drive, Calendar, Contacts, Photos or any other Google service, and we cannot read, modify or delete any content in your Google account.

Limited Use

In practice this means we do not transfer Google user data to third parties except as necessary to provide or improve the Service, to comply with applicable law, or as part of a merger or acquisition; we do not use it for advertising of any kind; we do not sell it; we do not use it to train any AI or machine learning model; and no human reads it except with your explicit consent, to investigate a security incident or abuse, or where required by law.

Withdrawing access

You can revoke our access at any time from your Google Account permissions page at myaccount.google.com/permissions. Revoking access stops future Google sign-in but does not by itself delete your OneFast account; to delete the account and the identifiers we hold, use the account settings in the console or contact privacy@onefast.ai. On deletion we remove the stored Google identifier along with the rest of your account data, subject to the retention rules in Section 11.

5. The Content of Your API Requests

This section is the one that matters most for an inference API, so we state it precisely.

What we do with request content

When you call our API, the body of your request — your prompt, messages, system instructions, files, images and any other content — is held in memory only for as long as is needed to authenticate the request, select a route, forward it to the chosen model provider, stream the response back to you, and count tokens for billing. It is then discarded. It is not written to our databases.

The exceptions, in full

The Service does contain a diagnostic capability to record request and response bodies, because some faults cannot be diagnosed without seeing the exact payload that produced them. We may enable it only:

• to investigate a specific fault that you have reported to us, at your request and for the period needed to reproduce and resolve it;
• to investigate suspected abuse, fraud, a security incident, or a breach of our Terms, where we have a legitimate interest in doing so and narrower means are not sufficient; or
• where we are required to do so by a binding legal obligation.

When enabled, recording is scoped as narrowly as the situation permits, each stored body is truncated, access is restricted to administrators, and the records are deleted when the investigation concludes and in any event within 30 days. We will tell you if we enable recording for your account for fault investigation. We do not enable it to build a corpus, to analyse usage patterns, or for any product-development purpose.

Transmission to model providers

We cannot answer your request without sending its content to a model provider. Depending on the model and route you select, that provider may be a model developer, a cloud platform hosting that developer’s model, or an authorised reseller. Once your content reaches them, their own privacy terms, retention periods, abuse-monitoring practices and training policies apply, and those vary between providers.

Some providers retain inputs for a period for abuse monitoring; some offer zero-retention handling; some may use inputs to improve their models unless a specific tier or setting is used. If a particular guarantee is material to you, select a model or route whose provider offers it, and review that provider’s terms before sending data. We are happy to discuss the arrangements applicable to a specific route — contact privacy@onefast.ai.

Prompt caching

Several models support prompt caching, in which the provider stores a processed form of a repeated portion of your prompt for a short period to reduce cost and latency. Where you use caching, that cached content resides with the model provider under its terms, not with us. We record only the number of cached tokens, for billing.

Training

We do not use your inputs or outputs to train, fine-tune or evaluate any model, and we do not sell or license them to anyone for that purpose.

6. Routes and Upstream Data Practices

The same model can often be reached through more than one commercial route — for example directly from the model developer, through a cloud platform hosting that model, or through an authorised distributor. Routes differ in price, capacity and, importantly, in the data terms the upstream operator offers. The model name alone therefore does not determine how your data is handled; the route does.

We group routes into tiers so that this is a choice you can make rather than one you discover after the fact:

• Verified routes — supplied by model developers, their cloud platforms or authorised distributors, and selected because the upstream terms provide for no retention of request content and no use of it for model training. These are the routes intended for confidential, regulated or personal data.
• Cost-optimised routes — selected primarily for price or regional capacity. Upstream data terms vary between operators: some retain request content for a period for abuse monitoring, and some may use it to improve their services. These routes are not appropriate for confidential, regulated or personal data unless we have confirmed the applicable terms to you in writing.

We will tell you, on request, which tier a given route belongs to, the category of upstream operator serving it, the retention and training position of that operator so far as it is disclosed to us, and the processing region where that information is available. Enterprise customers can have these recorded in a route schedule that forms part of their contract. Contact privacy@onefast.ai or your account contact.

Where we cannot obtain a satisfactory written data commitment from an upstream operator, we treat that route as unverified, exclude it from verified-route traffic, and do not represent it as offering retention or training guarantees.

7. How We Use Personal Data

We use personal data to:

• provide the Service: authenticate you, route and deliver requests, and operate your dashboard;
• meter and bill usage, maintain your balance, process top-ups and refunds, and produce your usage records;
• operate routing and capacity: apply rate limits and quotas, detect degraded upstream routes, and fail over between them;
• maintain security and prevent abuse: detect credential compromise, fraud, payment abuse, scraping and violations of our Terms;
• provide support and respond to your enquiries;
• improve the reliability and performance of the Service using aggregated metadata — never request content;
• send service communications such as balance alerts, security notices, incident notifications and changes to these documents; and
• comply with law, respond to lawful requests, and establish, exercise or defend legal claims.

We send marketing email only if you have opted in, and every such message includes an unsubscribe link. Service and security notices are not marketing and cannot be opted out of while you hold an account.

8. Legal Bases for Processing

Where the EU or UK General Data Protection Regulation applies to our processing, we rely on the following legal bases:

• Performance of a contract — to create and operate your account, route your requests, and bill you.
• Legitimate interests — to secure the Service, prevent fraud and abuse, maintain and improve reliability, understand at country level which markets we serve, and defend legal claims. We balance these interests against your rights and use the least intrusive means available.
• Legal obligation — to meet accounting, tax, sanctions-screening and lawful-request requirements.
• Consent — for optional marketing communications and non-essential cookies. You may withdraw consent at any time, without affecting processing carried out before withdrawal.

9. How We Share Personal Data

We do not sell personal data. We share it only as follows:

• Model providers — the content of a request and the metadata necessary to serve it, as described in Section 4.
• Payment processors — to take payment, issue refunds and manage disputes. We use third-party processors and receive only a customer reference and transaction status, not full card details.
• Identity providers — where you choose federated sign-in, in order to verify your identity with that provider.
• Infrastructure and operations providers — hosting, network, storage, email delivery and error-monitoring providers acting on our instructions under written terms.
• Reseller and organisation administrators — if your account was created under a reseller or organisation, that administrator can see your account identity, balance, and usage metadata for the accounts under their management. They cannot see your request content, because it is not recorded.
• Legal and safety recipients — where we reasonably believe disclosure is required by law or necessary to protect the rights, property or safety of any person, and where lawful and practicable we will notify you first.
• Corporate transactions — in a merger, acquisition, financing or sale of assets, subject to the acquirer honouring this policy.

10. International Data Transfers

We operate internationally and the Service is deliberately multi-region. Personal data may be processed in Hong Kong, mainland China, Singapore, the United States, the European Union and other locations where our infrastructure providers or the model provider serving your request operates. The physical location where your request is served depends on the route you select.

Where we transfer personal data out of the European Economic Area, the United Kingdom or another jurisdiction that restricts transfers, we rely on an adequacy decision where one applies, or otherwise on Standard Contractual Clauses or an equivalent lawful transfer mechanism agreed with the customer concerned. Cost-optimised routes may involve operators for which those contractual safeguards are not available; customers with regulated transfer requirements should use only routes identified as eligible in their route schedule. You may request details of the mechanism applicable to a specific transfer.

11. Retention

• Account data — retained while your account is open, and deleted or anonymised after closure except where we must keep it for legal, accounting or fraud-prevention purposes.
• Usage and billing records — retained for as long as needed for billing, dispute resolution, accounting and tax compliance. Detailed per-request records are subject to periodic cleanup once they are no longer needed for these purposes.
• Request and response bodies — not retained. Where recording is enabled under an exception in Section 4, retained no longer than the investigation requires and in any event no more than 30 days.
• Support correspondence — retained for as long as needed to handle the matter and to maintain a record of it.
• Web server and security records — retained for a short period for security monitoring and incident investigation.

Backups are retained on a rolling basis and are overwritten in the ordinary course. Data that has been deleted from the live systems may persist in a backup until that backup expires; it is not restored into live use.

12. Security

We protect the Service with measures appropriate to the risk, including: TLS on connections between you and our production gateway; storage of passwords as salted hashes; optional two-factor authentication; hardware-key authentication for privileged interfaces; scoped, quota-limited and revocable API keys with optional model restrictions and IP allowlisting; least-privilege administrative access with role separation; and network-level isolation of production systems.

Transport between our gateway and an upstream operator is route-specific. We require TLS on verified routes; we do not represent transport as uniformly encrypted across every route, and we will confirm the position for a specific route on request.

We describe here only the controls we operate. We do not claim encryption at rest, customer-managed keys, formal data residency, SOC 2, ISO 27001, HIPAA eligibility, PCI scope, or independent penetration testing, and you should not infer any of them from this policy. Where an enterprise engagement requires a specific control, we will state in the contract whether we have it.

The main practical risk to your data is a leaked API key. Do not embed keys in client-side code or public repositories, rotate them periodically, and revoke any key you suspect has been exposed.

No system is perfectly secure. If a breach affecting your personal data occurs, we will notify you and the relevant regulators where required, and within the timescales set by applicable law.

13. Cookies and Similar Technologies

We use cookies and local browser storage for two purposes:

• Strictly necessary — to keep you signed in, secure your session, and remember interface preferences such as language, theme and display currency. These cannot be disabled without breaking the Service.
• Analytics — where used, to understand aggregate usage of the site so we can improve it. Analytics cookies are optional and are set only with consent where consent is required.

We do not use advertising cookies and we do not permit third parties to track you across other sites through our Service. Our API endpoints do not set cookies. You can block or delete cookies through your browser settings; blocking strictly necessary cookies will prevent you from signing in.

14. Your Rights

Subject to applicable law, you have the right to access the personal data we hold about you; to receive it in a portable form; to have inaccurate data corrected; to have data deleted; to restrict or object to certain processing, including processing based on legitimate interests; and to withdraw consent where processing is based on consent.

Much of this is self-service: your account settings let you view and correct your profile, view and export your usage records, rotate or revoke API keys, and close your account. For anything else, contact privacy@onefast.ai. We will respond within one month, or sooner where the law requires, and may need to verify your identity first. We do not charge for these requests unless they are manifestly unfounded or excessive.

If you are in the EEA or the UK, you may lodge a complaint with your local data protection authority. We would appreciate the chance to address your concern first.

California residents

We do not sell or share personal information as those terms are defined under the California Consumer Privacy Act, and we have not done so in the preceding twelve months. California residents may exercise rights of access, deletion, correction and non-discrimination through the contact address above.

Mainland China residents

Where the Personal Information Protection Law applies, you additionally have the right to request an explanation of our processing rules and to have personal information transferred to another handler where technically feasible. Cross-border transfers described in Section 9 are made with your separate consent where that law requires it.

15. Children

The Service is not directed to anyone under 18 and we do not knowingly collect personal data from children. If you believe a child has provided us with personal data, contact privacy@onefast.ai and we will delete it.

16. Changes to This Policy

We may update this policy. Where a change materially affects how we handle your personal data — in particular any change to the commitments in Section 4 — we will give at least 30 days’ notice by email to the address on your account or by prominent notice in the Service before it takes effect. Other changes take effect when posted, and the “Last updated” date above will reflect them.

17. Contact

For privacy questions, requests to exercise your rights, or to request a data processing agreement, contact privacy@onefast.ai.

LinkOne AI Limited

Enterprise customers requiring a data processing agreement, a subprocessor list, standard contractual clauses or a route-specific data schedule should contact the same address; these are handled as contract documents rather than published here.