Skip to main content
Conversation history enables your AI Voice Agent to maintain context across multiple conversation turns.

​
What is Conversation History?

Conversation history is the record of all messages in a session:
  • User messages - All user inputs (text and voice)
  • Agent responses - All agent responses
  • Tool calls - Tool execution records
  • Tool results - Results from tool calls

​
How It Works

​
Automatic Management

History is automatically managed: 
# Server side - automatic
session.add_message(user_message)  # Add user message
session.add_message(agent_response)  # Add agent response
session.get_conversation_history()  # Get full history

​
Context for LLM

The agent uses conversation history as context: 
Conversation History
    ↓
LLM Processing
    ↓
Response Generation

​
History Contents

​
Message Types

History includes:
  • User messages - Text and voice transcripts
  • Agent responses - Text responses
  • Tool calls - When agent calls tools
  • Tool results - Results from tool execution

​
Example History

[
    {"role": "user", "content": "What's the weather in London?"},
    {"role": "assistant", "content": "The weather in London is cloudy, 15°C."},
    {"role": "user", "content": "What about tomorrow?"},
    {"role": "assistant", "content": "Tomorrow in London will be sunny, 18°C."}
]

​
Best Practices

​
History Length

  • Keep focused - Long conversations may exceed context limits
  • Use tools - Tools can help manage context
  • Clear when needed - Start new session for new topics

​
History Quality

  • Clear messages - Help the agent understand intent
  • Relevant tool results - Tools should return useful information
  • Consistent terminology - Use consistent terms

​
Next Steps