Your AI Connector Docs

What You Can and Cannot Send

Two sets of rules apply to every message you send through the platform: the platform’s, and the rules of the channel the message travels on. Both matter, and the channel’s rules are usually the stricter of the two.

The Platform’s Rules

You agree not to use the platform, and not to let your clients or end users use it, to:

  • Send unsolicited bulk messages or spam, or message people who have not opted in where opt-in is required.
  • Harass, threaten, defame, abuse, intimidate or stalk anyone.
  • Distribute illegal content, child sexual abuse material, content that infringes someone else’s rights, or content that incites violence or hatred.
  • Distribute malware, viruses, ransomware or phishing links.
  • Break any law that applies to your business or to the people you are messaging.
  • Misrepresent who a message is from, including impersonating the platform, its staff, or anyone else.
  • Hide from people that they are talking to an AI where the law that applies to your audience requires you to disclose it, or instruct the AI to deny being an AI or to claim to be human. There is a built-in AI transparency setting for exactly this; if you turn it off, meeting any disclosure obligation is on you.
  • Break the rules of the platforms the app connects to, including Meta’s Business Platform Terms, the WhatsApp Business Messaging Policy, Twilio’s messaging policies and Stripe’s acceptable use policy.
  • Send SMS or WhatsApp messages that break marketing communication law where your recipients are, for example the US Telephone Consumer Protection Act and A2P 10DLC rules, the European ePrivacy Directive, or Canada’s anti-spam law (CASL).
  • Reverse engineer the platform, or probe and test its security without written permission.

Suspected breaches are investigated, and access can be suspended or ended for repeated, severe or unresolved breaches. Where possible you are told first and given a chance to fix it.

The Channel’s Rules Also Apply

This is the part people underestimate. Every message you send on WhatsApp, Instagram, Facebook Messenger or SMS is also governed by that provider’s own policy, and they publish their own lists of restricted business categories and prohibited content. Those rules apply to your messages whether or not the platform’s own policy mentions them, and the provider decides what happens if you break them. The usual consequence is your number, page or account being restricted, and nobody here can appeal it on your behalf.

In practice there are two layers that can stop a message, and neither is a decision made by our support team:

  • The AI provider. The models that write your replies have their own content rules. If a conversation is about something they refuse (adult content, weapons, illegal activity and similar categories), the reply simply does not get generated. Nothing is sent, and no amount of instruction in your agent changes that.
  • The channel provider. Meta (WhatsApp, Instagram, Messenger), your SMS carrier and the other channels enforce their own restricted-business lists. This is the layer that hurts: they do not block a single message, they restrict or ban the whole number, page or account, and that is not something we can reverse for you.

So the honest answer to “can I use this for X” is usually not about our policy at all. If the AI provider will not write it, the agent goes quiet. If the channel provider does not allow it, you lose the channel. Check both before you build.

Two practical consequences:

  • Check the channel’s policy for your industry before you build. If your business is in a category messaging providers treat as sensitive or restricted, the channel’s rules decide whether you can message there at all.
  • Getting a number or account banned is the risk that actually bites. Almost every ban we see comes from messaging people who did not ask to hear from you. See Scaling WhatsApp Outreach Safely.

Countries and Sanctions

Where you and your clients can use the platform is decided by the channel providers and the payment processor more than by us:

  • WhatsApp Business API. Meta does not offer the WhatsApp Business Platform to businesses in Cuba, Iran, North Korea, Syria, or the Crimea, Donetsk and Luhansk regions, and phone numbers from those places (Syria is +963) can neither be connected nor receive messages sent through the Business API. That is Meta’s rule, enforced on their side, and it applies through every provider. Yemen (+967) is not on that list.
  • WhatsApp Web works wherever the normal WhatsApp app works, because it is your own WhatsApp account linked to the platform. Telegram, the website chat widget, email and custom channels carry no country restriction from us. Instagram and Messenger follow Meta’s own rules for business accounts in those countries.
  • Payments. Stripe, which handles billing, lists the same places (Cuba, Iran, North Korea, Syria, Crimea, Donetsk and Luhansk) as prohibited, so a business based there cannot pay directly. For an agency this usually does not matter: the agency is the customer and pays for the plan, and bills its own clients however it likes.
  • Our terms require you to comply with export and sanctions law. The service cannot be used from a country under comprehensive sanctions, or by anyone on a restricted-party list, and that applies to your clients as well.

Not Sure Whether Your Use Case Is Allowed?

Ask before you build it. Email support with a short description of who you would message, what about, and how they opted in, and you will get a straight answer on whether it is something we can support. That is a much better outcome than finding out after your number has been restricted.