Keep spam out, credits in, and customers safe
AI spam detection pauses junk chats before they burn credits, rate limits and blocklists stop floods, and actions only touch data for verified customers — after they confirm.
- AI spam detection
- Rate limit per visitor
- Blocked countries
What you get
AI spam detection
Chats are checked after the 2nd, 4th, 8th and 16th message. Ads, link farms, gibberish, bots and jailbreak attempts are paused, so they stop using credits — rude but real customers aren't. Describe spam in your own words.
Rate limit per visitor
Cap how many messages one visitor can send in a time window, with your own message when they go over.
Blocked countries
Visitors from countries you choose don't see the chat bubble or your help center.
Blocklists
IP addresses and ranges, emails or whole domains, and phone numbers. Their chats are paused right away.
Pause and block from the inbox
Pause the AI in any chat, block a visitor in one click, or mark it “not spam” to let the AI answer again.
Per-agent credit cap
Limit how many credits one agent can use in a month, with your own message when it's reached.
Verified customers for actions
Order and billing actions only work for customers who proved who they are.
Confirmation and approval
Changes need the customer's yes, and can wait for a team member's approval.
Allowed domains
Choose which websites may show your chat bubble.
FAQ
Questions people ask
Still wondering about something? Ask us or read the docs.