Skip to main content

​
Memory Management Strategies

Configure memory behavior for optimal performance and cost.

​
Limiting Message History

from upsonic import Agent, Chat

agent = Agent("openai/gpt-4o")

chat = Chat(
    session_id="session1",
    user_id="user1",
    agent=agent,
    num_last_messages=20  # Only keep last 20 messages in context
)

​
Memory with Summarization

from upsonic import Agent, Chat

agent = Agent("openai/gpt-4o")

chat = Chat(
    session_id="session1",
    user_id="user1",
    agent=agent,
    full_session_memory=True,
    summary_memory=True  # Automatically summarize long conversations
)

​
Tool Call Results in Memory

from upsonic import Agent, Chat

agent = Agent("openai/gpt-4o")

chat = Chat(
    session_id="session1",
    user_id="user1",
    agent=agent,
    feed_tool_call_results=True  # Include tool execution results in memory
)

​
User Profile Management

from upsonic import Agent, Chat

agent = Agent("openai/gpt-4o")

chat = Chat(
    session_id="session1",
    user_id="user1",
    agent=agent,
    user_analysis_memory=True,
    user_memory_mode="update"  # Incrementally update profiles
)

​
Dynamic Profile Schema

from upsonic import Agent, Chat

agent = Agent("openai/gpt-4o")

chat = Chat(
    session_id="session1",
    user_id="user1",
    agent=agent,
    dynamic_user_profile=True  # Generate profile schema automatically
)

​
Memory Optimization Tips

  1. Use num_last_messages to limit context size
  2. Enable summary_memory for long conversations
  3. Use user_memory_mode="update" for incremental profiles
  4. Disable feed_tool_call_results if tool outputs are large