should()

should() Terms of Service

These terms cover should(), the decision service run by HyperMindZ at should.hypermindz.ai: the web app, the HTTP API, the MCP server, the TypeScript client and the CLI. By requesting access you accept them (version v1.1). How we handle data is in the Privacy Policy.

1. Developer preview

should() is a developer preview. It may change, pause or stop without notice, and there is no service-level commitment. Access is by approval, and we may suspend an account that breaks these terms or puts the service at risk. It is free during the preview. If that changes, we will tell you before you are charged.

2. Your account and keys

  • Keep your API keys secret. Anything done with your keys or your sign-in is done by you. Revoke a key you think is exposed.
  • Connected sources (for example Google Drive) belong to you alone. You decide which keys and agents may read them, and you can disconnect at any time.

3. What you send us

  • You keep ownership of your questions, context, judgments and connected documents. You give us permission to process them to provide, secure and improve the service, including aggregate analysis by HyperMindZ staff, as the Privacy Policy describes. We don’t use them to train models.
  • Only send content you are allowed to share. Don’t send health records, payment card data, government ID numbers, or special-category personal data. The preview is not built for them.
  • Meeting transcripts and shared documents can contain other people’s words and details. You are responsible for having the right to use them.

4. Decisions are probabilistic

  • A should() answer is a model’s estimate, not advice and not a guarantee. ESCALATE means a person should decide. You are responsible for what your software or your team does with an answer.
  • Don’t let should() alone approve actions that spend money, change contracts or affect people’s rights without a review step you control.
  • Extracted fields are in beta and not yet measured for accuracy. Have a person review decisions that rely on them.

5. Third-party model providers: good-faith reliance

should() does not run its own models. Each decision is sent through OpenRouter to a model provider: today TypeSafe AI (Jev), and for some judgments, condensing and extraction, OpenAI models hosted by Microsoft Azure. The current list is in the Privacy Policy.

  • We chose these providers, and configured our account and requests, in good faith, based on the data-retention and training policies they publish. We ask for zero data retention and no training, and we route only to providers that say they meet those settings.
  • Those policies belong to the providers, not to us. We cannot inspect their systems or independently verify that they follow their published policies. OpenRouter itself says its list of zero-retention providers reflects what providers publish or represent, and is not OpenRouter’s guarantee.
  • The policies are not always consistent or complete. For example, a provider’s own terms may reserve rights (such as abuse monitoring or derived telemetry) that its listing elsewhere does not mention. Where we find a gap like this, we ask the provider in writing, and we say in the Privacy Policy what is still unconfirmed.
  • Providers can change their policies. We review them periodically. If we learn that a provider no longer meets our settings, we stop sending it data and tell you.
  • To the extent the law allows, HyperMindZ is not liable for a model provider’s handling of data that goes against the policies we relied on. Our own obligations (what we store, for how long, and who can see it) are ours, and are set out in the Privacy Policy.

6. Acceptable use

Don’t use should() to break the law, to probe or disrupt the service, to get around rate limits or spend caps, to extract others’ data, or to resell access. We rate-limit and cap usage to keep the preview fair and within budget.

7. Disclaimers and liability

The service is provided "as is" during the preview, without warranties of any kind, to the extent the law allows. HyperMindZ’s total liability for any claim about the preview is limited to the greater of the amount you paid us in the last 12 months or USD 100.

8. Changes

We may update these terms. The version and date are at the bottom of this page, and each access request records the version accepted. For a material change we will tell approved users before it takes effect.

9. Governing law

HyperMindZ is a Delaware corporation. These terms are governed by the laws of the State of Delaware, without regard to its conflict-of-laws rules. Any dispute about these terms or the service will be brought in the state or federal courts located in Delaware, and you and HyperMindZ consent to their jurisdiction.

10. Contact

Questions about these terms: contactus@hypermindz.ai.

Version v1.1 · last updated 2026-09-28