# Add\_messages annotation doesn't seem to be working while building basic chatbot **URL:** \ **Category:** LangGraph\ **Tags:** cloud, intro-to-langgraph, python-help\ **Created:** [July 21, 2025, 12:05pm UTC](https://forum.langchain.com/t/add-messages-annotation-doesnt-seem-to-be-working-while-building-basic-chatbot/597 "2025-07-21T12:05:11Z")\ **Posts on this page:** 2\ **Page:** 1
**Author:** ![amittendulkar](https://yyz1.discourse-cdn.com/flex007/user_avatar/forum.langchain.com/amittendulkar/32/283_2.png) [@amittendulkar](https://forum.langchain.com/u/amittendulkar)\ **Post date:** [July 21, 2025, 12:05pm UTC](https://forum.langchain.com/t/add-messages-annotation-doesnt-seem-to-be-working-while-building-basic-chatbot/597/1 "2025-07-21T12:05:12Z")
I have just started with langgraph and I am facing very basic issue. I am following the “Build a basic chatbot” page in the Getting started section (link: [1. Build a basic chatbot](https://langchain-ai.github.io/langgraph/tutorials/get-started/1-build-basic-chatbot/) It seems the add\_messages annotation is not working for me. The AI is not recognizing the context. My code, ```python import os from langchain_openai import AzureChatOpenAI from dotenv import load_dotenv _ = load_dotenv() model = AzureChatOpenAI( azure_endpoint=os.environ['AZURE_OPENAI_ENDPOINT'], azure_deployment=os.environ['AZURE_OPENAI_DEPLOYMENT_NAME'], openai_api_version=os.environ['AZURE_OPENAI_API_VERSION'], ) from typing import Annotated from typing_extensions import TypedDict from langgraph.graph import StateGraph, START, END from langgraph.graph.message import add_messages class State(TypedDict): messages: Annotated[list, add_messages] graph_builder = StateGraph(State) def chatbot(state: State) -> State: print('state', state) updated_state = {'messages': [model.invoke(state['messages'])]} print('updated_state', updated_state) return updated_state graph_builder.add_node('chatbot', chatbot) graph_builder.add_edge(START, 'chatbot') graph_builder.add_edge('chatbot', END) graph = graph_builder.compile() def stream_graph_updates(user_input: str): for event in graph.stream({'messages': [{'role': 'user', 'content': user_input}]}): for value in event.values(): print('Assistant:', value['messages'][-1].content) while True: try: user_input = input('User: ') if user_input.lower() in ['quit', 'exit', 'q']: print('Goodbye!'); break stream_graph_updates(user_input) except: user_input = 'What do yoy know about LangGraph?' print('User: ' + user_input) stream_graph_updates(user_input) break ``` Here is the interaction, ```auto User: What is 2+2? state {'messages': [HumanMessage(content='What is 2+2?', additional_kwargs={}, response_metadata={}, id='20178aec-155f-4081-b505-3ec5545ec17f')]} updated_state {'messages': [AIMessage(content='2 + 2 equals **4**! 😊', additional_kwargs={'refusal': None}, response_metadata={'token_usage': {'completion_tokens': 11, 'prompt_tokens': 14, 'total_tokens': 25, 'completion_tokens_details': {'accepted_prediction_tokens': 0, 'audio_tokens': 0, 'reasoning_tokens': 0, 'rejected_prediction_tokens': 0}, 'prompt_tokens_details': {'audio_tokens': 0, 'cached_tokens': 0}}, 'model_name': 'gpt-4o-2024-11-20', 'system_fingerprint': 'fp_ee1d74bde0', 'id': 'chatcmpl-BvjSCXYRhPQJ9jl5ylNS5O9GDLv0n', 'service_tier': None, 'prompt_filter_results': [{'prompt_index': 0, 'content_filter_results': {'hate': {'filtered': False, 'severity': 'safe'}, 'jailbreak': {'filtered': False, 'detected': False}, 'self_harm': {'filtered': False, 'severity': 'safe'}, 'sexual': {'filtered': False, 'severity': 'safe'}, 'violence': {'filtered': False, 'severity': 'safe'}}}], 'finish_reason': 'stop', 'logprobs': None, 'content_filter_results': {'hate': {'filtered': False, 'severity': 'safe'}, 'protected_material_code': {'filtered': False, 'detected': False}, 'protected_material_text': {'filtered': False, 'detected': False}, 'self_harm': {'filtered': False, 'severity': 'safe'}, 'sexual': {'filtered': False, 'severity': 'safe'}, 'violence': {'filtered': False, 'severity': 'safe'}}}, id='run--c33d943d-f9e8-4c07-bab0-fd951cf61ceb-0', usage_metadata={'input_tokens': 14, 'output_tokens': 11, 'total_tokens': 25, 'input_token_details': {'audio': 0, 'cache_read': 0}, 'output_token_details': {'audio': 0, 'reasoning': 0}})]} Assistant: 2 + 2 equals **4**! 😊 User: What if I use subtraction in my earlier message? state {'messages': [HumanMessage(content='What if I use subtraction in my earlier message?', additional_kwargs={}, response_metadata={}, id='19dcbb5b-3d72-4a59-b98a-8b4db07be424')]} updated_state {'messages': [AIMessage(content='It seems like you\'re asking what would happen if you use subtraction in a message or a question. Could you clarify what you mean by "my earlier message"? Are you referring to a specific calculation or context? If you\'re thinking about using subtraction to express or solve something, feel free to elaborate so I can assist you further!', additional_kwargs={'refusal': None}, response_metadata={'token_usage': {'completion_tokens': 65, 'prompt_tokens': 17, 'total_tokens': 82, 'completion_tokens_details': {'accepted_prediction_tokens': 0, 'audio_tokens': 0, 'reasoning_tokens': 0, 'rejected_prediction_tokens': 0}, 'prompt_tokens_details': {'audio_tokens': 0, 'cached_tokens': 0}}, 'model_name': 'gpt-4o-2024-11-20', 'system_fingerprint': 'fp_ee1d74bde0', 'id': 'chatcmpl-BvjSeh3Rn4XwkeHp1jbBkFyJp8NFR', 'service_tier': None, 'prompt_filter_results': [{'prompt_index': 0, 'content_filter_results': {'hate': {'filtered': False, 'severity': 'safe'}, 'jailbreak': {'filtered': False, 'detected': False}, 'self_harm': {'filtered': False, 'severity': 'safe'}, 'sexual': {'filtered': False, 'severity': 'safe'}, 'violence': {'filtered': False, 'severity': 'safe'}}}], 'finish_reason': 'stop', 'logprobs': None, 'content_filter_results': {'hate': {'filtered': False, 'severity': 'safe'}, 'protected_material_code': {'filtered': False, 'detected': False}, 'protected_material_text': {'filtered': False, 'detected': False}, 'self_harm': {'filtered': False, 'severity': 'safe'}, 'sexual': {'filtered': False, 'severity': 'safe'}, 'violence': {'filtered': False, 'severity': 'safe'}}}, id='run--3be5e449-9fe4-4ad1-b940-9db6ff1cef07-0', usage_metadata={'input_tokens': 17, 'output_tokens': 65, 'total_tokens': 82, 'input_token_details': {'audio': 0, 'cache_read': 0}, 'output_token_details': {'audio': 0, 'reasoning': 0}})]} Assistant: It seems like you're asking what would happen if you use subtraction in a message or a question. Could you clarify what you mean by "my earlier message"? Are you referring to a specific calculation or context? If you're thinking about using subtraction to express or solve something, feel free to elaborate so I can assist you further! User: q Goodbye! ``` My python version, Python 3.11.2 Showing portion of `pip list` command, ```auto langchain 0.3.26 langchain-community 0.3.27 langchain-core 0.3.68 langchain-groq 0.2.4 langchain-openai 0.3.28 langchain-text-splitters 0.3.8 langgraph 0.5.3 langgraph-api 0.2.98 langgraph-checkpoint 2.1.0 langgraph-cli 0.3.5 langgraph-prebuilt 0.5.2 langgraph-runtime-inmem 0.6.0 langgraph-sdk 0.1.72 langsmith 0.4.8 ``` Does it have to do with my model with older version in the `.env` file? ```auto AZURE_OPENAI_API_VERSION=2024-12-01-preview ``` ---
**Author:** ![niilooy](https://yyz1.discourse-cdn.com/flex007/user_avatar/forum.langchain.com/niilooy/32/2886_2.png) [@niilooy](https://forum.langchain.com/u/niilooy)\ **Post date:** [July 21, 2025, 12:52pm UTC](https://forum.langchain.com/t/add-messages-annotation-doesnt-seem-to-be-working-while-building-basic-chatbot/597/2 "2025-07-21T12:52:22Z")
The loop restarts the graph fresh each time, so context won’t persist across multiple inputs. Following the next steps ([3. Add memory](https://langchain-ai.github.io/langgraph/tutorials/get-started/3-add-memory/)) you can compile the graph with **MemorySaver()** for state persistence across turns using a consistent **thread\_id** in stream calls.