Is there a built-in debugger or visual trace tool for Google ADK Agents like LangGraph’s debugger?

Gemini API
,
1

I’m trying to run an Explicit Invocation workflow using an AgentTool in Google ADK. However, it’s not working as expected; my Planner Agent produces a FunctionCall, but the tool doesn’t always run correctly, or the input schema doesn’t match. Is there any official or recommended way to debug ADK multi-agent workflows step by step? For example, how can I see the InvocationContext, session state, function call arguments, and agent handoffs in detail, similar to how LangGraph provides a visual debugger for graph execution?

3

Hey @Sohaib_Sajid ,

Great question! As of now, Google ADK doesn’t offer a built-in visual debugger like LangGraph. For step-by-step debugging, you might need to log the InvocationContext , session state, and function calls manually. Hopefully, a more visual or integrated debugging tool is on the roadmap soon!

Thanks!