More control over your AI agent conversations

We’ve added three new security features to Guardrails, giving you more control over who can access your AI agent and how unwanted conversations are handled.

Spam detection Automatically detect and pause conversations based on criteria you define. Add a custom prompt describing what your agent should consider spam, and Chatbase will check the user’s 2nd, 4th, 8th, and 16th messages. If the content matches your criteria, the conversation is paused automatically.

Pause conversations manually You can now pause any conversation directly from Chat logs, giving you a quick way to stop a conversation whenever needed.

Country blocking Block visitors from selected countries from accessing your AI agent. Country blocking applies to the website chat bubble, Center Stage, Help Page, and API, without affecting WhatsApp, Messenger, or other integration channels.

Together, these new controls give you more flexibility to protect your agent from abuse and manage who can interact with it.