Terms of service
The agreement covering Tartarus AI accounts: what a plan buys, acceptable use, payment and refunds, and how either side ends it. Last updated 6 August 2026.
1. The agreement
These terms cover your use of Tartarus AI, operated by Tartarus AI. Creating an account means you accept them. You must be old enough to enter a contract where you live.
2. Your account
One account belongs to one person. You are responsible for keeping your password and API tokens private, and for everything done with them. Tell us promptly if either is exposed; a token can be revoked from the account area at any time.
3. Plans and payment
Plans are prepaid in cryptocurrency for a fixed period and activate once the payment confirms on chain. Nothing renews automatically. Prices are shown in US dollars and the amount payable in a given currency is fixed at the rate quoted when the payment is created.
All payments are final. Because a plan is prepaid and access begins the moment the payment confirms on chain, confirmed payments are non-refundable — in full or in part, used or unused, and including any time left on the period if you cancel or stop using the service before it ends. The only exception is a refund right the law gives you that we cannot contract away.
4. Limits
Each plan carries weekly and per-session token budgets, a context ceiling and a per-response output cap, published on the pricing page. Reaching a limit pauses generation until the budget resets; it does not affect access to your history. We may adjust published limits with notice to account holders.
5. Acceptable use
Tartarus AI runs models with reduced refusal training, and the point of that is to stop legitimate work being blocked for resembling something else. It is not an absence of rules. This section is the complete list, and it is short deliberately: everything not on it is your business rather than ours.
Prohibited outright
- Any material that sexualises a minor, in any form, including fiction and any depiction generated or described. This line does not move and there is no context in which we will discuss it.
- Direct assistance in attacking, intruding into or disrupting computer systems you neither own nor have written authorisation to test. Security research on your own systems, or on systems you are contracted to assess, is expressly permitted.
- Using the service to defraud, stalk, threaten, harass or impersonate a real person, or to produce material intended to cause them harm.
- Serious violence against people: operational assistance in producing weapons capable of mass casualties, or in planning an attack on people or infrastructure.
- Anything illegal where you are. Output generated by a model does not change the law that applies to producing or possessing it.
Account rules
Do not resell access, share credentials, or attempt to circumvent plan limits or rate limits. One account belongs to one person.
Expressly permitted
Stated because the assumption tends to run the other way in this category: fiction containing violence, cruelty, addiction, crime or morally repugnant characters; adult creative writing about adults; security research and malware analysis you are authorised to perform; frank discussion of health, legal, financial, political and historical subjects; and research into extremism, propaganda, abuse and drug policy. None of these require permission from us.
Automated traffic is welcome through the API on plans that include it, within the published rate limits. Traffic that degrades the service for others may be throttled.
6. Output
You own what you send and, as between you and us, what the model writes back in response. Model output can be wrong, out of date or confidently misleading, including when sources are cited. It is not professional advice. Check anything you intend to rely on.
7. Availability
We aim for continuous availability but do not guarantee it. Maintenance, upstream failures and capacity limits can interrupt the service. The service is provided as is, without warranties beyond those that cannot be excluded by law.
8. Liability
To the extent the law allows, our total liability arising from your use of Tartarus AI is limited to the amount you paid us in the twelve months before the claim, and we are not liable for indirect or consequential loss.
9. Ending it
You can stop using Tartarus AI and delete your account at any time. We may suspend or close an account that breaches section 5, or where we are required to by law. Because a plan is prepaid and, per section 3, all payments are final, closing an account does not refund any time remaining on it.
10. Changes
We may update these terms. Material changes are announced by email to account holders before they take effect, and the date at the top of this page changes with them. Continuing to use the service after that date means you accept the revised terms.
This document is a starting template, not legal advice. Have a lawyer in your operating jurisdiction review it before launch, particularly sections 3 and 8.