Terms of Service

These Terms of Service (the Terms) are an agreement between you and Real Fake Intelligence, Inc., a Delaware corporation (RFI, we, us), covering your use of the rfi.dev website, the RFI account portal, and the RFI inference API (together, the Service).

By creating an account, buying credits, or sending a request to the API, you accept these Terms. If you are accepting on behalf of a company or other organisation, you represent that you have authority to bind it, and you means that organisation.

How we handle personal data is described in our Privacy Policy, which forms part of these Terms.

You must be at least 18 years old to use the Service. The Service is a paid developer tool and is not directed at children. If we learn that an account holder is under 18, we will close the account and refund any unspent credit balance.

You may not use the Service if you are barred from doing so under applicable sanctions or export-control law, or if we have previously terminated your account.

Where we operate. We are a United States company, we operate the Service from the United States, and we direct it at users there. We do not currently market the Service in the European Economic Area or the United Kingdom, and we do not offer the local representation or transfer machinery that serving those markets properly would require. Nothing stops you signing up from outside the United States, and we do not block you — but if you do, you are choosing a US-operated service, the data we keep is stored in the United States while inference may run on rented servers in other countries (see our Privacy Policy), and you are responsible for whether that works for your own legal obligations. If you need us to serve an EEA or UK market properly, email support@rfi.dev: we would rather hear it as a request than guess.

You sign in through a third-party identity provider (currently Google or GitHub). You are responsible for keeping access to that identity secure.

API keys are bearer credentials. Anyone holding a key can spend your credits. You are responsible for all activity under your keys, including usage you did not intend, and for revoking a key promptly if it is exposed. You can create and revoke keys at any time in the portal. We will never ask you for a key over email or chat.

You must give us an email address that reaches you. Billing and service notices are sent there.

• Prepaid only. You buy credits before you use the Service. There is no invoicing and no post-paid billing: you are never charged for usage after the fact, only for credits before it. When your balance reaches zero, further API requests are refused until you top up.
• If your balance goes below zero. We refuse new requests once your balance reaches zero, but a request already in flight can complete afterwards and leave the balance slightly negative. You owe us that amount. We may take it out of your next top-up, and we may suspend the account until it is cleared.
• Automatic top-up, if you turn it on. You can ask us to top your balance up automatically. If you do, you authorise us to charge the payment method you have saved with Stripe, each time your balance falls below the threshold you set, for the amount needed to bring your balance up to the target you set — so the amount varies with how far below the threshold your balance has fallen. It is never less than the gap between your threshold and your target, and it is subject to the daily limits on your account. This is not a subscription: there is no periodic charge and no renewal term — a charge happens only when your own usage takes you below your threshold. We email a receipt for every top-up. You can turn automatic top-up off at any time in the portal; it takes effect immediately except for a charge already in progress. If a charge fails we will email you and may switch automatic top-up off.
• Credits do not expire. An unspent balance stays on your account for as long as the account is open.
• Deleting your account forfeits your balance. Credits do not expire while your account is open, but if you delete your account, any unspent balance is forfeited and is not refunded. If you simply want to stop using the Service, you do not need to delete the account — leave it open and the credits stay. Deleting the account also does not clear a negative balance.
• Credits are not refundable except where a refund is required by applicable law. They have no cash value, cannot be transferred between accounts, and cannot be redeemed for anything other than use of the Service.
• What a request costs is set by the rate card published on the Service at the time of the request, metered per token. We record the token counts and the charge for every request, and you can see them in the portal.
• Requests we shed are not billed. If we refuse a request because of capacity — an HTTP 429 or 529, or any other refusal we make to protect capacity — you are not charged for it.
• Prices can change. We will publish a change to the rate card before it takes effect. A price change never applies retroactively to usage already metered, and never changes the dollar value of credit already on your balance.
• Payments are processed by Stripe. Card details are entered on Stripe's hosted checkout and are never received or stored by RFI. Your purchase is also subject to Stripe's terms.
• Taxes. Prices are exclusive of tax; where we are required to collect it, we will add it at checkout. Any other taxes are your responsibility where they apply, except for taxes on our own income.
• Chargebacks. If you dispute a charge with your card issuer rather than with us, we may suspend the account and freeze the balance while the dispute is open. We would much rather you email us first.

You may not use the Service to:

• break the law, or help someone else break it, in any jurisdiction that applies to you or to us;
• generate or distribute child sexual abuse material, non-consensual intimate imagery, or content that sexualises minors;
• harass, threaten, defame, or impersonate a real person, or generate content presented as a real person's words without their consent;
• build weapons, or produce operational instructions for attacks on people, infrastructure, or computer systems;
• run large-scale disinformation, spam, or fraud, including generating fake reviews, fake identities, or content that impersonates an organisation;
• attack the Service itself — bypassing rate limits or authentication, probing for vulnerabilities without our written permission, reselling access in a way designed to obscure who is calling us, or extracting model weights;
• make automated decisions with legal or similarly significant effects on a person (credit, employment, housing, insurance, or the administration of justice) without meaningful human review.

Outputs are generated, not verified. The model produces plausible text, not established fact, and it will sometimes be confidently wrong. You are responsible for reviewing output before you rely on it, and for any decision you make on the strength of it. Do not use the Service as the sole basis for medical, legal, or financial advice to another person.

If you are building a product on top of the Service, you are responsible for your own users' conduct and for telling them, in substance, what this section says.

You keep your inputs. We claim no ownership of the prompts, files, or other content you send (Input). You grant us a non-exclusive, worldwide, royalty-free licence to receive, process, transmit and cache your Input — including in the shared prefix cache described in our Privacy Policy — and to generate, inspect and return Output, for these purposes and no others: (a) providing the Service to you; (b) the operational purposes described in our Privacy Policy — keeping the Service working, secure and correctly billed, and investigating a security incident, a suspected breach of section 5 (Acceptable use), or a support request you have made; (c) computing aggregate service-quality measures — error rates, including the rate at which tool calls come back malformed — from requests and responses as they pass through, which means reading enough of them in memory to count what failed and keeping the count rather than the content; and (d) complying with law. Where we act as your processor, those purposes are your documented instructions to us.

Outputs are yours. As between you and us, we assign to you all right, title and interest we may have in the content the model returns (Output). To the extent that assignment is not effective for any reason — including because no such rights exist — we grant you a perpetual, irrevocable, worldwide, royalty-free, non-exclusive licence to use, reproduce, modify and distribute the Output for any purpose. This assignment and licence are subject to Output is not exclusive below, and to the licence under which the underlying model is made available to us: where that licence imposes terms on the use of Output, those terms apply to you too, and we link them from our documentation.

You warrant that your Input does not infringe anyone's rights and that you have the rights you need to send it.

We do not train on your data. We do not use your Input or Output to train or fine-tune any model, or to evaluate one, except on a request you have asked us to look into. We do not sell it or share it for anyone else's training. We serve an open-weights model that we did not train; nothing you send is fed back into it.

Output is not exclusive. The Service is used by many customers and the model is deterministic given the same inputs and settings, so another customer may receive output identical or similar to yours. We make no claim that your Output is unique, and no representation about whether Output is copyrightable in your jurisdiction.

Feedback you send us about the Service — bug reports, feature requests, benchmarks — we may use freely and without obligation to you.

Where we act as your processor for Input, we will: process it only on your documented instructions, which are these Terms and our Privacy Policy; keep it confidential and require the same of anyone who handles it for us; apply the security measures described in the Privacy Policy; use only the categories of vendor described there, name the specific vendors in each category on request, and update that section before we add a category; give you reasonable help with data-subject requests and with your own security, breach-notification and impact-assessment obligations, at your cost where the help is substantial; tell you without undue delay if we become aware of a personal-data breach affecting your Input; and delete Input at the end of the Service, subject to the retention periods in the Privacy Policy.

If you need a standalone Data Processing Agreement — including the named list of our subprocessors, which we do not publish in the Privacy Policy — email support@rfi.dev and we will send one; it replaces this section. We do not offer EU-specific transfer terms today; see section 2 on where we operate.

We do not offer a service-level agreement, an uptime guarantee, or service credits. The Service is provided on a reasonable-efforts basis. We may throttle, shed, or queue requests to protect capacity, and we may change, suspend, or discontinue any part of the Service.

If we permanently discontinue the Service, we will give you notice at the email on your account and refund your unspent credit balance. That is one of a small number of refunds we commit to; the others are in sections 2, 9 and 13.

Model versions change. We may add, retire, or update the models we serve; where a change would break existing integrations we will try to give advance notice, but we do not guarantee that a specific model version remains available.

You can stop using the Service at any time and delete your account from the portal.

We may suspend or terminate your access if you breach these Terms, if your usage threatens the security, integrity, or availability of the Service or another customer's use of it, or if we are required to by law. Where the circumstances allow it, we will tell you why and give you a chance to fix it first; where they do not — an active attack, a legal demand, a serious acceptable-use breach — we may act immediately.

On termination, your right to use the Service ends. If we terminate you for convenience, we refund your unspent balance. If we terminate you for a breach of section 5, we may decline to.

Sections 4, 6, 7, 10, 11, 12, 14, 15 and 16 survive termination, along with any other provision that by its nature is meant to.

The Service is provided “as is” and “as available”, without warranty of any kind. To the fullest extent permitted by law, we disclaim all warranties, express or implied, including merchantability, fitness for a particular purpose, non-infringement, and any warranty arising out of a course of dealing or usage of trade.

We do not warrant that the Service will be uninterrupted, timely, secure, or error-free, or that Output will be accurate, complete, current, or fit for your purpose. Some jurisdictions do not allow certain disclaimers, in which case this section applies to the extent they do.

To the fullest extent permitted by law, neither party will be liable for indirect, incidental, special, consequential, exemplary, or punitive damages, or for lost profits, lost revenue, lost data, or business interruption, even if advised of the possibility.

Our total aggregate liability arising out of or relating to these Terms or the Service will not exceed the greater of one hundred US dollars ($100) or the total amount you paid us in the twelve months before the event giving rise to the claim.

The exclusions and the cap above do not apply to your obligations under section 12, to your obligation to pay amounts you owe us, or to any liability that cannot be excluded or limited under the law that applies — which in some places includes liability for fraud and for deliberate wrongdoing. Some jurisdictions do not allow these limits, in which case they apply to the extent they do.

You will defend and indemnify us against third-party claims, and the reasonable costs and damages that come with them, arising out of your Input, your use of Output, your breach of these Terms, or your violation of law or of a third party's rights. We will notify you of the claim promptly — a delay reduces your obligation only to the extent it actually prejudices the defence — and you may control the defence with counsel of your choice. We may take part with our own counsel at our own expense. You may not settle a claim in a way that admits fault on our part, imposes any obligation or payment on us, or restricts how we operate the Service, without our prior written consent, which we will not unreasonably withhold. This section does not apply to the extent the claim arises from our own breach of these Terms.

We may update these Terms. If a change is material, we will give at least 30 days' notice — by email to the address on your account, or a notice in the portal — before it takes effect. The date at the top of this page always reflects the current version.

Continuing to use the Service after a change takes effect means you accept it. If you do not accept it, stop using the Service before the effective date and email us; we will refund your unspent balance.

Talk to us first, and it is a condition of starting anything else. Before starting an arbitration or a lawsuit, you must email support@rfi.dev with (a) your name and the email address on your account, (b) a description of the dispute, and (c) the relief you want. We will send you the same before we start anything. Neither of us may begin a proceeding until 30 days after that notice is sent, and any limitation period on the claim is paused while those 30 days run. A notice covers one claimant and one account; one notice cannot cover several people.

Binding arbitration. Except as stated below, any dispute arising out of or relating to these Terms or the Service will be resolved by binding individual arbitration administered by JAMS under its Streamlined Arbitration Rules, before one arbitrator, in English. Where you are using the Service for personal, family, or household purposes, JAMS' Consumer Arbitration Minimum Standards also apply and prevail over anything inconsistent in this section. The arbitration will be conducted on documents and by video conference unless the arbitrator orders a hearing in person; any in-person hearing will be held in the county where you live or work, or in Delaware, at your election. Judgment on the award may be entered in any court of competent jurisdiction.

Who pays. We will pay all JAMS filing, administrative and arbitrator fees above what it would have cost you to file the same claim in court, to the extent the applicable JAMS rules and minimum standards require. Each of us pays our own legal costs, unless a statute or the arbitrator's award says otherwise.

Class-action waiver. Disputes will be arbitrated only on an individual basis. You and RFI each waive any right to bring or participate in a class, collective, consolidated, or representative action, and the arbitrator may not consolidate more than one person's claims. If this waiver is found unenforceable as to a particular claim, that claim — and only that claim — is severed from this section and proceeds in court under section 15; the rest of this section stays in force for every other claim.

This section is not to be rewritten. Section 16's severability provision does not permit this section to be read as allowing class, collective, consolidated, or representative arbitration. If such arbitration is nevertheless ordered as to any claim, this section is void as to that claim and section 15 governs it.

Public injunctive relief. If the law that applies to you gives you a right to seek public injunctive relief that cannot be waived or sent to arbitration, that claim — and only that claim — is severed from this section and stayed until the arbitration of your individual claims is finished. Nothing else in this section is affected.

Many similar claims at once. If 25 or more claimants represented by the same or by coordinated counsel bring demands raising substantially similar claims, those demands will be administered under JAMS' mass-arbitration or equivalent procedures in effect at the time, and the limitation period for every claimant in the group is paused while that process runs.

Who decides what. The arbitrator decides all issues, including the scope, interpretation and enforceability of this section — except that a court, not an arbitrator, decides whether the class-action waiver and the public-injunctive-relief paragraph above are enforceable.

You can opt out, and it costs you nothing. Email support@rfi.dev with the subject “arbitration opt-out” and the email address on your account, within 30 days of the date you first accept these Terms, which is the date your account was created and which we record. We will confirm your opt-out by reply email; keep that reply. Opting out has no effect on anything else in this agreement and we will not treat you differently for it. If you opt out, disputes go to the courts named in section 15.

Exceptions. Either of us may bring an individual claim in small-claims court in any place where that court has jurisdiction over the other of us; section 15's exclusive-venue provision does not apply to such a claim. Either of us may also seek injunctive relief in court for infringement or misuse of intellectual property or for unauthorised access to the Service.

These Terms are governed by the laws of the State of Delaware, without regard to its conflict-of-laws rules, and by the Federal Arbitration Act as to section 14. For any dispute not subject to arbitration, you and RFI submit to the exclusive jurisdiction of the state and federal courts located in Delaware.

If you are a consumer resident in a jurisdiction whose law grants you rights that cannot be waived by contract, nothing here removes them.

• Entire agreement. These Terms and the Privacy Policy are the whole agreement between us about the Service and supersede anything said before.
• One year to bring a claim. Any claim arising out of or relating to these Terms or the Service must be brought within one year after it arises, or it is permanently barred — unless the law that applies to you does not allow that.
• Legal demands for your data. If we receive a legal demand for your data, we will tell you before we respond unless we are legally prohibited or it would risk someone's safety, and we will push back on demands that are overbroad or improper.
• Aggregate and de-identified data. We may create and use aggregated or de-identified information derived from your use of the Service — for example, usage volumes, latency and error rates — to operate, secure, measure and improve the Service, and to publish statistics that do not identify you or any customer.
• Severability. If a provision is unenforceable, it is limited to the minimum extent necessary and the rest stays in force.
• No waiver. Not enforcing a provision once does not waive it.
• Assignment. You may not assign these Terms without our written consent. We may assign them to an affiliate or in connection with a merger, acquisition, or sale of assets.
• Force majeure. Neither party is liable for a delay or failure caused by something outside its reasonable control.
• No third-party beneficiaries. These Terms create rights only between you and us, with one exception: our officers, directors, employees, contractors and affiliates may rely on sections 10 and 11 as if they were named in them.
• Notices. We reach you at the email on your account, and a notice we send there is effective when sent, whether or not you read it — so keep the address current. You reach us at support@rfi.dev.
• Relationship. Nothing here creates a partnership, agency, or employment relationship.

Real Fake Intelligence, Inc.
Email: support@rfi.dev