Not seeing LLM thought process/thinking steps despite enabling stream_intermediate_steps

I’ve successfully set up agent-ui with my reasoning agent, but I’m not seeing the detailed LLM thought process/thinking steps that are shown in the README examples and documentation.

Current Configuration:

reasoning_agent = Agent(
    name="Reasoning Agent",
    model=OpenAIChat(id="gpt-4o"),
    tools=[
        ReasoningTools(add_instructions=True),
        YFinanceTools(stock_price=True, analyst_recommendations=True, company_info=True, company_news=True), 
    ],
    instructions="Use tables to display data.",
    markdown=True,
    stream=True,
    stream_intermediate_steps=True,
)

What I’m Seeing:

  • :white_check_mark: Streaming responses work correctly
  • :white_check_mark: Tool calls are displayed and executed
  • :white_check_mark: REASONING section shows steps like “STEP 1 - Analyze Semiconductor Companies”
  • :cross_mark: Missing: Detailed thinking process like “PLANNING SEMICONDUCTOR MARKET ANALYSIS” with explanations

Expected Behavior (from README):
The interface should display detailed thinking steps similar to the README screenshot showing:

  • “THINK” button/section
  • Detailed planning phases like “1. PLANNING SEMICONDUCTOR MARKET ANALYSIS”
  • Explanatory text: “The user wants a comprehensive analysis of four major semiconductor companies.”

Screenshots:

Current Output: [Image 1 - shows my current interface]

Expected Output: [Image 2 - shows README example with detailed thinking]
I can not provide it here because it lets new users one image embedding.

Questions:

  1. Is there an additional configuration needed to enable the detailed thinking display?
  2. Does this require a specific model or reasoning mode?
  3. Is the ReasoningTools(add_instructions=True) sufficient, or are there other parameters needed?
  4. Could this be related to the OpenAI model not providing intermediate reasoning steps?

Additional Context:

  • My agent works perfectly for tool calling and final responses
  • All streaming functionality is working as expected
  • I’m specifically missing the internal thought process visualization shown in Agno examples

Related Configuration:
If there are specific agent-ui frontend settings or backend configurations needed to display thinking steps, please let me know what I might be missing.

Hey @burakkaanozbek , thanks for reaching out and supporting Agno. I’ve shared this with the team, we’re working through all requests one by one and will get back to you soon.
If it’s urgent, please let us know. We appreciate your patience!