Terms of service

LikelyAI estimates the likelihood that submitted text, images, or audio were generated or materially modified by AI. It returns a probability, a separate confidence in that probability, a classification, and the individual signals behind it.

These terms cover the API, the dashboard, the documentation, and this website. By creating an account you agree to them. The operator is LikelyAI, legal entity name (to be confirmed).

This is the most important clause here, so it is not buried at the bottom. LikelyAI produces a probabilistic estimate. It does not prove authorship, it cannot prove authorship, and no output should be described to anyone as proof.

A result can be wrong in both directions. Human writing can score high. AI output can score low. Absence of provenance metadata is not evidence that something was made by a person. Weak, short, corrupted, or heavily transformed input lowers the confidence score rather than the probability, and an inconclusive classification means we are declining to estimate.

You therefore agree that you will not use a LikelyAI result as the sole basis for a decision with significant consequences for a person. That includes, without limitation, academic penalties, disciplinary action, refusing or terminating employment, denying access to a service, or any legal claim. Where a result informs such a decision, a human must review it alongside other evidence, and the person affected should be able to respond.

• You must be old enough to enter a contract where you live, which is generally 18.
• You are responsible for everything done with your API keys. A key is shown once, so store it somewhere safe, and revoke it in the dashboard if it may have leaked.
• You are responsible for what members of your organisation do. Roles and permissions are yours to manage.
• Do not share an account, and do not resell access without a written agreement with us.

Do not use the service to:

• Submit content you have no right to submit.
• Harass, profile, or target an individual, or build a system whose purpose is to do so.
• Present our output as a definitive verdict, or misrepresent what a probability and a confidence score mean.
• Attempt to circumvent rate limits, spending limits, or authentication.
• Reverse engineer the models, or use outputs to train a competing detector.
• Break applicable law, including data protection law, in the jurisdiction you operate in.

We may suspend access if use of the service threatens its stability, security, or legality.

• Usage is measured in credits, not requests, because a short comment and a long recording are not the same amount of work. Credit weights are configuration and every billing record stores the pricing version applied, so past invoices stay reproducible if weights change.
• A plan includes a monthly credit allowance. Included credits are consumed first, and usage beyond the allowance is billed as overage at your plan’s rate.
• You can set usage alerts, a soft limit, and a hard limit. A hard limit causes further requests to be rejected rather than billed, and the API returns a specific error code for it.
• Subscription fees are billed in advance and overage in arrears. Fees are exclusive of tax unless stated otherwise.
• Rate limits apply per plan. Exceeding them returns an error rather than queueing your requests indefinitely.

Current prices are on the pricing page and are still being calibrated. Notice period for a price change, and the refund and cancellation policy, are to be confirmed (to be confirmed).

You keep all rights to the content you submit. You grant us only the narrow, temporary licence needed to process it and return a result, and that licence ends when the content is deleted. We do not use your content to train our models. See the privacy policy for retention detail.

We keep all rights to the service, the models, and the software. The detection results we return are yours to use within these terms.

We aim to keep the service available and will publish status information, but the Developer and Growth plans are provided without a service level guarantee. A contractual SLA is available on the Scale plan. We may change or deprecate parts of the API, and when we do we will version the change and give reasonable notice before removing anything you depend on.

To the extent the law allows, the service is provided as is, without warranty of accuracy, fitness for a particular purpose, or uninterrupted availability. Given the nature of the product, we specifically make no warranty that any individual result is correct.

We are not liable for indirect or consequential loss, and our total liability is capped at cap, typically the fees paid in the preceding 12 months (to be confirmed). Nothing here excludes liability that cannot lawfully be excluded.

You can close your organisation at any time from the dashboard. We may suspend or terminate access for a material breach of these terms, for non-payment, or where required by law, and we will tell you why unless we are legally prevented from doing so. On termination your keys stop working and your data is deleted on our normal schedule, except billing records we are required to retain.

We may update these terms. For material changes we will give notice before they take effect, and the date at the top reflects the current version. Continuing to use the service after that date means you accept the updated terms.

Governing law and the forum for disputes are to be confirmed with counsel (to be confirmed).