Chathistory

📸 Screenshots

Here are visual examples of this section:

Chathistory - Worker Configuration Interface

The Chathistory worker manages conversation history by storing and retrieving chat messages from a database. It supports different history modes including full history retention and automatic summarization when conversations become too long. The worker can maintain conversation context while optimizing storage and token usage through intelligent summarization.

2. Configuration Parameters

history: Sets the history mode - can be "full" for complete history or "sumarized" for automatic summarization
keepLatest: Number of recent messages to keep unsummarized (default: 100)
sumarizeWhen: Threshold number of messages that triggers summarization (default: 200)
sumarizePrompt: Custom prompt used for summarizing chat history (default: system prompt)
sumarizationModel: AI model used for summarization (default: "openai/gpt-4-turbo")

3. Input/Output Handles

output: Output handle - returns array of chat history items including messages, roles, and content

4. Usage Examples with Code

// The worker automatically retrieves history based on user ID and agent ID
// History items are returned with the following structure:
{
  role: "user" | "assistant" | "system",
  type: "message",
  content: [{ text: "message content", type: "text" }],
  __FROM_DB__: true // indicates item was loaded from database
}

5. Integration Examples

This worker integrates at the beginning of chat workflows to provide conversation context to other workers. It works seamlessly with chat completion workers and can be combined with search and retrieval workers for enhanced context management.

6. Best Practices

Set appropriate keepLatest values based on your token limits and context requirements
Use summarization mode for long-running conversations to manage costs
Customize the summarization prompt to preserve domain-specific information
Monitor token usage when using full history mode with lengthy conversations

7. Troubleshooting Tips

Ensure user ID is provided in agent parameters to avoid "No uuid provided" errors
Check database connectivity if history retrieval fails
Verify API keys are configured properly when using summarization features
Review summarization model availability if summarization fails