How can I display the execution steps of the Llama_index AI Agent?

Other Questions AI Agent
1

Hey everyone, I’m working on an AI Agent project with the Llama_index Python library and I need help figuring out how to display its execution steps. I first tried enabling the verbose mode by setting verbose=True in the FunctionAgent constructor, but it didn’t work:

from llama_index.core.agent.workflow import FunctionAgent

agent_instance = FunctionAgent(
    tools=[sample_tool],
    llm=sample_llm,
    system_prompt='Act as a helpful assistant',
    verbose=True,
    allow_parallel_tool_calls=True
)

Then, I attempted to use callbacks:

from llama_index.core.callbacks import CallbackManager, LlamaDebugHandler

debug_handler = LlamaDebugHandler()
callback_manager = CallbackManager([debug_handler])

agent_instance = FunctionAgent(
    tools=[sample_tool],
    llm=sample_llm,
    system_prompt='Act as a supportive assistant for engagement queries',
    callback_manager=callback_manager,
    allow_parallel_tool_calls=True
)

Unfortunately, neither method seems to work. I’m currently using Llama_index version 0.12.31. Has anyone encountered this issue before or has any ideas on how to display the steps properly?

2

hey samuel87, i ran into a similar problem. try using the LlamamindHandler instead of LlamaDebugHandler. it worked for me:

from llama_index.core.callbacks import CallbackManager, LlamamindHandler

llamamind_handler = LlamamindHandler()
callback_manager = CallbackManager([llamamind_handler])

hope this helps! let me know if u need more info