Skip to the document
Tidal AI

Version 2026-10-08 · effective [[TO FILL: effective date]]

Acceptable Use Policy

What you, and the bots acting for you, may not do with Tidal AI. It applies to everything in your workspace, including what bots do on their own within the permissions you give them.

The short version

Use Tidal AI lawfully, honestly and safely. Your bots act for you, so what they do counts as what you do. Don't use them to harm, deceive, spam or attack anyone, and don't try to get around the service's limits.

Not allowed

Illegal or harmful activity

  • Anything illegal where you are or where it has effect, or that helps someone else break the law.
  • Content or actions that sexually exploit or endanger children, in any form. We report it to the authorities.
  • Violence, threats, harassment, stalking, doxxing, or inciting hatred against people or groups.
  • Developing or obtaining weapons capable of mass harm (biological, chemical, nuclear, radiological), or attacks on critical infrastructure.
  • Malware, ransomware, phishing kits, or code meant to damage, disable or get unauthorised access to systems.
  • Fraud, scams, impersonation, fake reviews, or deceiving people about who or what they are dealing with.
  • Infringing others' intellectual property, privacy or other rights, including collecting personal data without a lawful basis.

Email and messaging

  • No unsolicited bulk or marketing email. Bots may email people who expect to hear from you (customers, colleagues, people who wrote first, people who asked to be contacted). They may not send cold marketing campaigns, newsletters to bought or scraped lists, or any bulk messaging. Per-domain daily limits enforce this.
  • Don't remove, hide or alter the AI-disclosure footer, the X-AI-Generated header or the unsubscribe link on bot email, or keep emailing someone who unsubscribed (it is blocked automatically).
  • Don't use bots to send messages that pretend to be written by a human when a person asks, or that impersonate a real person or organization.

The service itself

  • Don't attack, probe, overload or disrupt Tidal AI or its providers, or test their security, except under our [[TO FILL: security disclosure / bug bounty policy link]].
  • Don't get around rate limits, budgets, approvals, sanctions blocks, or account restrictions, or create accounts to avoid a suspension.
  • Don't use your bots' machines for crypto mining, scanning networks, sending spam, hosting phishing pages, or other abuse of compute or networking.
  • Don't try to reach other customers' data or our internal systems.
  • No raw model access. Don't use Tidal AI to resell, sublicense or expose access to the underlying AI models (for example, a public completions API or an "LLM gateway"), or to build a competing model service.
  • No jailbreaking or red-teaming of AI models through Tidal AI (attempts to make a model ignore its safety rules), unless we have agreed it in writing.
  • Don't use output to train models that compete with the model providers', where their terms forbid it.

High-risk decisions

  • Don't use Tidal AI to make decisions with legal or similarly significant effects on people (credit, employment, housing, insurance, education, legal status, access to essential services) without meaningful human review.
  • Don't use it for medical diagnosis or treatment, or legal or financial advice to others, without qualified human review.
  • Don't use it for biometric identification of people, social scoring, or other practices prohibited by the EU AI Act.

Elections and public information

  • No deceptive content about elections or voting, or content that impersonates candidates or officials.
  • Don't present AI-generated images, audio or video of real people as authentic.

Model provider policies

Tidal AI uses AI models from several providers, and their usage policies apply to your use of their models through us. By using a model you agree to its provider's policy, including: [[TO FILL: links to the current usage policies of Anthropic, OpenAI, Google, Meta (Llama), Mistral, xAI, DeepSeek and the other model makers offered]]. Where a provider's policy is stricter than this one, the stricter rule applies.

How we enforce it

We use automated checks (rate limits, per-domain email limits, abuse signals) and review reports. If we find a problem we may remove content, pause a bot or mission, block a feature, suspend or close an account, and report illegal activity to the authorities. Where we can, we tell you what happened and give you a chance to respond.

Report abuse

If you see abuse of Tidal AI (including an email from a bot that you didn't want), tell us at [[TO FILL: abuse contact email]]. Every bot email also has a one-click unsubscribe link.