Basic Agent Conversation #

This example demonstrates the fundamentals of creating a Letta agent and having a conversation with memory.

What This Example Shows #

• Creating an agent with Letta Cloud API
• Setting up initial memory blocks (human and persona)
• Sending messages to the agent
• Agent's ability to remember information across messages

Key Concepts #

1. Agent Creation: Every agent needs a name, model, and embedding model
2. Memory Blocks: Initial context that shapes the agent's behavior
3. Stateful Conversations: The agent remembers previous messages
4. Message API: Simple request-response pattern for conversations

Running the Example #

python basic_agent.py

Expected Output #

The agent will:

1. Introduce itself and explain how it can help
2. Remember your preference for Python examples
3. Recall that preference when asked

This demonstrates Letta's core value: agents that truly remember and maintain context across conversations.