🤖 AG2 + GitHub CopilotKit + FastAgency

[elCaptnCode]

🤖 AG2 + GitHub CopilotKit + FastAgency — run_group_chat Workflow

📝 Context

Following the AG2 CopilotKit Integration blog post,
I’m running run_group_chat(pattern, starting_message) and aiming to pass the response into ui.process(response) just like in the example.

pattern = AutoPattern(
        initial_agent=sales_pitch_coordinator,
        agents=[
            sales_pitch_coordinator,
            interviewer_agent,
        ],
        user_agent=user,
        group_manager_args={
            "llm_config": llm_config,
        },
    )

    response = run_group_chat(
        pattern=pattern,
        messages="Hello! Let's get started by interviewing the user about the current situation.",
    )

    return ui.process(response)

---

🛑 Issue

When I try this, the response isn’t coming back as the expected IOMessage type — which means ui.process() won’t accept it.

[ERROR] Message is not an IOMessage: type='group_chat_run_chat' content=GroupChatRunChatEvent(uuid=UUID('8bf0917a-67b8-4a04-8170-48ff88436ab2'), speaker='Sales_Pitch_Coordinator', verbose=True)

---

❓ Questions

• Has anyone wired up run_group_chat with FastAgency + CopilotKit successfully?
• If so, how did you handle the returned event object vs. an IOMessage?
• Is there a recommended pattern for adapting the output before sending to the UI?

---

💬 Call for Input

Any tips, patterns, or working examples would be super helpful.
I’d love to understand if this is a workflow mismatch or if the output just needs to be adapted.