Memory & Context

Meebly agents maintain conversation context across sessions using thread-based memory and customer-scoped tracking. They remember your customers and the history of their interactions, enabling personalized and coherent cross-session conversations.

Thread-Based Memory

Every conversation is stored in a thread. When you start a chat, a unique threadId is returned in the complete event. Include this threadId in subsequent requests to continue the conversation.

// First message
POST /v1/chat
{
  "agentId": "agent_abc123",
  "environmentId": "env_xyz789",
  "messages": [{"content": "Hello"}]
}

// Complete event returns threadId
{
  "type": "complete",
  "data": {
    "text": "Hi! How can I help you?",
    "threadId": "thread_def456"
  }
}

// Continue conversation with threadId
POST /v1/chat
{
  "agentId": "agent_abc123",
  "environmentId": "env_xyz789",
  "threadId": "thread_def456",
  "messages": [{"content": "I need help with my order"}]
}

Continuing Conversations

The agent remembers the full conversation history when you provide a threadId. This enables:

• Follow-up questions without repeating context
• Multi-turn problem solving
• Personalized responses based on conversation history
• Maintaining context across page refreshes

Customer-Scoped Memory

Use the customerId parameter to track which user is chatting. This enables:

• Analytics per user
• User-specific conversation history
• Personalized agent behavior
• Usage tracking and billing

{
  "agentId": "agent_abc123",
  "environmentId": "env_xyz789",
  "messages": [{"content": "What's my order status?"}],
  "customerId": "customer_789",  // Your user identifier
  "threadId": "thread_def456"
}

Best Practices