Spaces

What is a trigger?

A trigger is something that happens, such as a customer sending a message, that wakes your AI agent so it can respond straight away.

In plain words

A shop doorbell is a trigger. It does not do the serving. It simply tells you someone has walked in, so you can look up and say hello.

In Spaces, triggers are the moments worth reacting to: a message arrives, a comment is left, someone follows you, an account needs attention, a chat is handed to a person. Each one reaches your agent as a small note saying what happened and where.

What your agent does next is up to you. You can tell it in plain words, for example: "When a customer asks about opening hours, answer from my notes, and pass anything about a refund to me."

A trigger plus a plain instruction is most of what you need.

A trigger only tells the agent what happened. It does not give the agent extra powers. Anything it sends still goes through the same limits and pacing as every other message.

How it works in Spaces

Spaces has a catalog of triggers, shown on the Triggers page, such as message.received, comment.received, message.sent, account.status and account.resting. Agents read them from the event stream, and software can receive them as webhooks.

Event names are exact, and each trigger carries the account, the chat or post, and the facts about what happened. Text is cut to a thousand characters, and payloads hold no tokens or file links.

Related terms

FAQ

Do I need to set triggers up one by one?

Usually not. You tell your agent what you want in plain words, and it reads the events it needs.

Does a trigger send a message by itself?

No. It only tells your agent something happened. Anything sent follows the usual limits.

What if my agent is off when something happens?

Spaces keeps recent events for 14 days, so it can catch up when it is back.

Where can I see the full list?

On the Triggers page, grouped by what they are about.

Useful next: All triggers, Features, API reference. Every term: the glossary.