Terms of service
These terms govern your use of AI Inference. By creating an account or sending a request through the platform, you agree to them.
1. What the service is
AI Inference is managed inference routing and metering. You send a request; we resolve it to a model in our catalog — the one you named, or, if you named none, the lowest-cost model you have enabled that offers the capability you asked for — dispatch the call, count the input and output tokens it reports, and price those tokens at the rate published for that model.
We are not the author of the models and we do not warrant their output. We make no representation about the accuracy, completeness, safety, legality or fitness for any purpose of anything a model returns, and nothing a model returns is advice of any kind. You are responsible for reviewing output before you rely on it or pass it to anyone else.
Each upstream provider operates under its own data policy, which is surfaced in the model detail view. Where you choose a model, you accept that policy for the traffic you send to it.
2. Your account and your keys
- Give accurate account information and keep your password to yourself.
- API keys are issued from your own console and the full key is displayed once, at the moment it is created. We store only a hash of it, so neither we nor our support can recover a key afterwards. If you lose one, delete it and issue another.
- You are responsible for everything done through your account and through any key issued from it. Tell us at omnichannel200@gmail.com if you believe either has been used by someone else.
- You must be old enough to enter a contract where you live, and authorised to act for any company you name.
3. Fees and payment
Usage is billed per token. Input and output are counted separately on every call and priced at the rate published for the model that served it; that rate is shown against each model in the catalog, in the product, before you send a request to it. There is no per-seat charge and no tier you are moved into.
- Cache hits are not billed. An identical prompt, to the same model, at the same temperature, inside the cache window set on your account, is served from your own cache at no charge.
- A free allowance applies to new accounts. The allowance is a platform setting, and the current figure is shown on your usage page. It is not a contractual entitlement: we may change it, and the figure shown in the product at any time is the one that applies.
- One usage record is written per request, under an idempotency key, so a retried request is not counted twice.
Payment is taken by our payment processor. Card details go to it directly and never reach our servers or our database. Taxes, where they apply, are yours.
Refunds are not automatic and there is no stated refund window. If something has gone wrong, write to omnichannel200@gmail.com and we will deal with it case by case. Usage already recorded remains payable; you may stop sending requests at any time, and usage recorded before you stopped will still appear on your next invoice.
4. Your prompts and your records
Your prompts, and the responses returned to you, are yours. We claim no ownership of them and we do not train on them. You grant us only the permission we need to route the request, meter it, cache it for you, and show you your own history.
Your records are scoped to your account, and the prompt cache is keyed per account, so one account is never served another account's response. What we retain about each call, and for how long, is set out in the privacy policy.
You are responsible for the content of what you send, for having the right to send it, and for any personal data you put into a prompt.
5. Acceptable use
You agree not to use AI Inference to:
- break the law, infringe someone else's rights, or generate material that sexually exploits a child;
- generate or distribute malware, credential-stuffing material, or content intended to defraud, impersonate or deceive a specific person;
- work around a rate limit, an access control or a billing control, including by opening additional accounts to obtain the free allowance more than once;
- attempt to reach another account's requests, cache entries, keys or usage, or to probe or load-test the platform without written permission;
- infer, reconstruct or publish our internal commercial arrangements, or use the platform principally to benchmark it for a competing service;
- resell or sublicense access without a licence agreement with us.
We may suspend an account we have reason to believe is breaching this section, and where we do we will tell you why.
6. Availability
We work to keep the platform available and will give notice of planned maintenance where we reasonably can. We do not offer an uptime commitment or a service-level agreement, and the service is provided as is, without warranties of any kind beyond those that cannot lawfully be excluded.
A request can be refused, and in normal operation is refused in these cases:
- the account has reached the requests-per-minute limit set on it;
- inference is paused for maintenance;
- the account has been suspended;
- no enabled model offers the capability the request asked for;
- the named model is not enabled, or the upstream call fails or times out.
We hold no security certification or audit attestation and claim none. If you need our security documentation for a diligence process, ask and we will send what we have.
7. Limitation of liability
To the fullest extent the law allows, neither party is liable to the other for indirect, incidental, special, consequential or punitive damages, or for lost profits, lost revenue, lost data or lost business opportunity, however caused.
Our total liability arising out of or relating to these terms is limited to the amount you paid us in the twelve months before the event giving rise to the claim. Nothing here limits liability for fraud, for death or personal injury caused by negligence, or for anything else that cannot lawfully be limited.
8. Suspension and termination
You may stop sending requests at any time. Deleting an account is not self-service today — ask us at omnichannel200@gmail.com and we will delete the account and its records. We may suspend or terminate an account that breaches these terms, that is being used unlawfully, or whose payment is reversed. Where we suspend an account we will tell you why and, where it is reasonable to do so, give you a chance to put it right.
There is no self-service export either. Ask us and we will send you a copy of your own records, and we will keep them available for a reasonable period after termination so that you can. Sections 4, 7 and 9 survive termination.
9. Changes, law and contact
We may change these terms. If a change materially reduces your rights we will give notice before it takes effect, and continuing to use the platform after that date means you accept the new version. The date at the top of this page is the current revision.
These terms are governed by the law of the jurisdiction in which AI Inference is established, and the courts of that jurisdiction have exclusive jurisdiction over any dispute, without affecting any right you have to bring a claim where you live.
Questions about these terms: omnichannel200@gmail.com. See also our privacy policy.