​

get_memories

​

Use Cases

• Load workspace context at the start of a conversation
• Review all stored memories
• Implement memory browsing interfaces
• Audit stored information

​

Parameters

​

Response

​

Example Request

{
  "limit": 20,
  "offset": 0
}

​

Example Response

{
  "success": true,
  "memories": [
    {
      "memory_id": "550e8400-e29b-41d4-a716-446655440000",
      "content": "User prefers dark mode UI",
      "created_at": "2025-02-02T12:00:00Z",
      "updated_at": "2025-02-02T12:00:00Z",
      "agent_id": "agent_123",
      "agent_name": "Code Assistant",
      "api_key_id": "key_456"
    }
  ],
  "total": 1,
  "limit": 20,
  "offset": 0
}

​

Pagination

// First page
{ "limit": 50, "offset": 0 }

// Second page
{ "limit": 50, "offset": 50 }

// Third page
{ "limit": 50, "offset": 100 }

​

Behavior

• Returns memories in reverse chronological order (newest first)
• Only returns memories for the project associated with the API key
• Empty result sets return an empty array, not an error

​

Error Handling

​

Best Practices

1. Use reasonable limits: Start with 50-100 memories per request
2. Implement pagination: For workspaces with many memories
3. Cache results: Consider caching memories that don’t change frequently
4. Filter client-side: If you need to filter by agent or date, do it after retrieval