Docs

English

简体中文

English

简体中文

Appearance

Chat Trigger Node ​

The Chat Trigger node triggers workflow execution through chat conversations. This is the core node for building conversational AI Agents. Each workflow can only have one Chat Trigger.

Use Cases ​

  • Intelligent Customer Service - User messages trigger customer service response flows
  • Conversational Assistants - Build ChatGPT-like conversational assistants
  • Q&A Systems - Trigger answer generation based on user questions
  • Task Execution - Trigger tasks through natural language commands
  • Information Collection - Conversational forms, collect info through chat
  • Interactive Queries - Natural language database or API queries

Node Features ​

Basic Characteristics ​

  • Unique Per Workflow - One workflow can only have one Chat Trigger
  • Workflow Entry - Starting point for conversational workflows
  • Real-time Interaction - Supports real-time user-AI dialogue
  • Session Persistence - Automatically maintains conversation context and session state
  • Custom Variables - Can collect additional user-provided information

Built-in Output Fields ​

javascript
$('Chat Trigger').message             // User message content
$('Chat Trigger').userId              // User ID
$('Chat Trigger').sessionId           // Session ID
$('Chat Trigger').timestamp           // Message timestamp
$('Chat Trigger').conversationHistory // Conversation history

Node Configuration ​

Basic Settings ​

Custom Variables (customVariables) ​

Define custom variables to collect additional information.

javascript
customVariables: [
  {
    label: "User Name",
    variable: "userName",
    type: "string",
    required: true,
    maxLength: 50
  },
  {
    label: "Age",
    variable: "age",
    type: "number",
    required: false
  }
]

// Access custom variables
$('Chat Trigger').userName
$('Chat Trigger').age

Variable Types: string, numberVariable Properties: label, variable, type, required, maxLength, hidden

Workflow Examples ​

Example 1: Simple Q&A Bot ​

Chat Trigger
  → LLM Node
    System Prompt: "You are a friendly customer service assistant."
    User Prompts: [$('Chat Trigger').message]
  → Answer Node

User: "Hello"
AI: "Hi! How can I help you?"

Example 2: Context-aware Conversation ​

Chat Trigger
  → LLM Node
    User Prompts: $('Chat Trigger').conversationHistory
  → Answer Node

User: "I want to buy a laptop"
AI: "What's your budget?"
User: "Around 8000"  // AI remembers context
AI: "For 8000 yuan, I recommend..."

Example 3: AI Agent Dialogue ​

Chat Trigger
  → AI Agent Node
    System Prompt: "You are an intelligent assistant."
    Tools: [Weather API, Order API, Calculator]
    Conversation History: $('Chat Trigger').conversationHistory
  → Answer Node

User: "What's the weather in Beijing?"
AI: (Calls weather API) "Beijing is sunny, 15°C."

User: "Where's my order?"
AI: (Calls order API) "Your order shipped, arrives tomorrow."

Example 4: Multi-turn Information Collection ​

Chat Trigger
  → AI Agent Node
    System Prompt: "Collect: name, phone, date, room type. Ask step by step."
    Tools: [Entity Recognition Tool]
  → Code Node (Check completeness)
  → Conditional Branch
    → [complete] → Create booking
    → [incomplete] → Ask for missing info

Best Practices ​

1. Design Reasonable Custom Variables ​

javascript
// Good - necessary variables only
customVariables: [
  {label: "Member ID", variable: "memberId", type: "string", required: true}
]

// Bad - too many variables
customVariables: [
  // 10+ variables... poor UX
]

2. Utilize Conversation History ​

javascript
// Good - with context
User Prompts: $('Chat Trigger').conversationHistory

// Bad - loses context
User Prompts: [$('Chat Trigger').message]

3. User Identity Recognition ​

javascript
HTTP Request Node
  URL: `https://api.example.com/users/${$('Chat Trigger').userId}`

LLM Node
  System Prompt: `User info: ${$('HTTP Request').body}
Provide personalized service.`

FAQ ​

Q: Chat Trigger vs Webhook Trigger? ​

A:

  • Chat Trigger: Conversational interaction, auto session management, max 1 per workflow
  • Webhook Trigger: API integration, no session concept, multiple allowed

Q: How long is conversation history saved? ​

A: Within session, all history accessible. After session ends, may be cleared. Save to database if long-term storage needed:

javascript
HTTP Request Node
  URL: "https://api.example.com/conversation-logs"
  Body: {
    sessionId: $('Chat Trigger').sessionId,
    history: $('Chat Trigger').conversationHistory
  }

Q: How to implement multi-turn dialogue? ​

A: Use conversationHistory:

javascript
LLM Node
  User Prompts: $('Chat Trigger').conversationHistory
// Includes all history, AI auto understands context

Q: How to test Chat Trigger workflows? ​

A:

  • Use "Run" button in node toolbar
  • Input test messages
  • View execution results

Next Steps ​