Skip to content

Chat trigger

The Chat trigger lets you talk to a workflow from a chat window in the editor. Each message you send starts a run, and the workflow's AI reply appears in the chat.

The chat dialog opened with Run Workflow on a workflow with a Chat trigger: two questions about resetting a VPN password, the AI replies under each, and the message box at the bottom.The chat dialog opened with Run Workflow on a workflow with a Chat trigger: two questions about resetting a VPN password, the AI replies under each, and the message box at the bottom.

When to use it

  • Building and testing an AI assistant, such as an HR helper or a knowledge-base Q&A, before you connect it to a real channel.
  • Trying multi-turn prompts against an AI Agent with conversation history.

The chat window lives in the Flero editor. There is no hosted chat page or embeddable widget. To let people chat with a workflow in Slack, use the Slack connector's real-time triggers instead (limited availability).

For one-shot AI tasks (summarise this article, classify this email), use a webhook, schedule or manual trigger with a single AI Agent node.

Starting a chat

  1. Add a Chat trigger and connect it to an AI Agent node.
  2. Save the workflow.
  3. Click the chat button beside the Chat node (it replaces the usual play button), or pick the Chat trigger from the editor's Run menu.
  4. Type a message and send it.

Configuration

Field Notes
Placeholder Hint text in the message box. Defaults to "Type a message...".
Maximum Length Characters allowed per message.
Memory Key Names the conversation. Messages are stored per key. Defaults to default. Change it to keep separate conversations.
Multiline Input Allow line breaks in a message.
Clear on Send Empty the message box after sending.

Trigger data

The Chat node outputs:

{
  "message": "What's our refund policy?",
  "userMessage": "What's our refund policy?",
  "timestamp": "2026-10-13T10:00:00.000Z",
  "length": 25,
  "conversationHistory": [
    { "role": "user", "content": "Hi" },
    { "role": "assistant", "content": "Hello. How can I help?" }
  ],
  "hasHistory": true
}

conversationHistory holds the earlier turns for the current memory key, without the message you just sent. It is missing on the first message of a conversation.

Pass {{ $node["Chat"].output.message }} to the AI Agent's prompt. Use your Chat node's name.

The reply

The chat shows the first node output in the run that has a content field. The AI Agent node produces one, so a Chat → AI Agent workflow works without extra setup. When no node produces content, the chat says "Execution completed but no AI response found."

Conversation memory

The conversation is saved with the workflow, under the node's memory key. Reopening the chat loads it again. To clear it, start a new conversation from the chat window and confirm with Start Fresh.

The AI Agent node also has its own memory settings. See AI Agent.

Found something out of date? This page lives in the Flero docs content set.