Skip to content

Conversation

@andy0liang
Copy link

This PR details what integrating with Lucidic AI's Agent Observability platform looks like for TheAgenticBrowser.

Main Changes:

  • lai.init() at the beginning of Orchestrator.run()
  • lai.create_step() at the beginning of the step loop
  • lai.create_event() and lai.end_event() around Pydantic calls (will be supported natively soon)
  • The call to analyze the screenshot with ImageAnalyzer is natively captured since it uses the OpenAI SDK (no need to manually create event)
  • lai.end_step() at the end of the step loop
  • lai.reset() at the end of Orchestrator.run() -> This is equivalent to calling lai.end_session() + resetting for the next session

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant