microsoft/autogen · error · ValueError

Assistant not initialized

Error message

Assistant not initialized

What it means

The _get_assistant_id property on OpenAIAssistantAgent throws ValueError until the OpenAI Assistant object has been created/loaded. Initialization is lazy (create_assistant/get_assistant and thread creation happen on first use, e.g. inside on_messages or an explicit ensure/init call). Accessing assistant-dependent members before that completes hits the None guard.

Source

Thrown at python/packages/autogen-ext/src/autogen_ext/agents/openai/_openai_assistant_agent.py:375

        """The types of messages that the assistant agent produces."""
        return (TextMessage,)

    @property
    def threads(self) -> AsyncThreads:
        return self._client.beta.threads

    @property
    def runs(self) -> AsyncRuns:
        return self._client.beta.threads.runs

    @property
    def messages(self) -> AsyncMessages:
        return self._client.beta.threads.messages

    @property
    def _get_assistant_id(self) -> str:
        if self._assistant is None:
            raise ValueError("Assistant not initialized")
        return self._assistant.id

    @property
    def _thread_id(self) -> str:
        if self._thread is None:
            raise ValueError("Thread not initialized")
        return self._thread.id

    async def _execute_tool_call(self, tool_call: FunctionCall, cancellation_token: CancellationToken) -> str:
        """Execute a tool call and return the result."""
        if not self._original_tools:
            raise ValueError("No tools are available.")
        tool = next((t for t in self._original_tools if t.name == tool_call.name), None)
        if tool is None:
            raise ValueError(f"The tool '{tool_call.name}' is not available.")
        arguments = json.loads(tool_call.arguments)
        result = await tool.run_json(arguments, cancellation_token, call_id=tool_call.id)
        return tool.return_value_as_string(result)

View on GitHub (pinned to 027ecf0a37)

Solutions

1. Trigger lazy initialization before touching assistant state: call await agent.on_messages(...) or the agent's documented init/ensure method (e.g. _ensure_initialized) first.
2. If on_messages itself raises this, inspect for an earlier swallowed exception during create_assistant (network, invalid API key) and fix the root cause.
3. Avoid relying on private members (_get_assistant_id); use the public API surface which initializes on demand.

Example fix

# before
agent = OpenAIAssistantAgent(name="a", instructions="...", model="gpt-4o",
                            client=cl, assistant_id="asst_123")
assistant_id = agent._get_assistant_id  # ValueError: Assistant not initialized

# after
resp = await agent.on_messages(
    [TextMessage(source="user", content="hi")], CancellationToken()
)  # lazy init completed inside on_messages
assistant_id = agent._get_assistant_id

resp = await agent.on_messages([TextMessage(source="user", content="init")], ct)
# _assistant is now set; assistant-dependent members are safe to use

When it happens

Trigger: Constructing OpenAIAssistantAgent(..., assistant_id="asst_...") and immediately reading agent._get_assistant_id or otherwise touching _assistant-dependent state before any on_messages/on_messages_stream call; an async init task not awaited; API errors during lazy init leaving _assistant None.

Common situations: Inspecting/testing agent internals right after construction; calling helper methods (upload, runs) that assume initialization without triggering it; initialization that failed silently due to a network/API error so later property access fails instead.

Related errors

• No response was generated
• Incorrect client passed to OpenAIAssistantAgent. Please use
• Unsupported tool type: {type(tool)}
• No tools are available.
• The tool '{tool_call.name}' is not available.

AI-assisted analysis of microsoft/autogen@027ecf0a37 (2026-08-15). Data as JSON: /api/errors/17eaa5e10a7bd197. Report an issue: GitHub.