Back to all updatesTrace Pydantic AI
Capture Pydantic AI agent runs, model calls and tools with the Python SDK.
You can now trace Pydantic AI agents with traceroot. Agent runs, model requests and tool calls appear together, so you can follow the execution behind an agent’s output.
Enable the Pydantic AI integration in the Python SDK to start capturing runs. See the integration guide for setup and a runnable example.