Asked 1 month ago by CelestialAstronaut041
How can I fix the 'unexpected keyword argument proxies' error in ChatOpenAI on Streamlit?
The post content has been automatically edited by the Moderator Agent for consistency and clarity.
Want to use over 200+ MCP servers inside your coding tools like Cursor?
I'm encountering a pydantic ValidationError when running my Streamlit app on the cloud. On localhost everything works as expected, but after a recent code update (and even after reverting to the previous working code) I now get the following error:
BASH────────────────────── Traceback (most recent call last) ─────────────────────── /home/adminuser/venv/lib/python3.12/site-packages/streamlit/runtime/scriptru nner/exec_code.py:88 in exec_func_with_error_handling /home/adminuser/venv/lib/python3.12/site-packages/streamlit/runtime/scriptru nner/script_runner.py:579 in code_to_exec /mount/src/langgraph_agent/app/app_st.py:212 in <module> 209 210 211 if __name__ == "__main__": ❱ 212 │ main() 213 .... .... .... /mount/src/langgraph_agent/agents/agents.py:40 in get_llm 37 │ 38 │ def get_llm(self, json_model=True): 39 │ │ if self.server == 'openai': ❱ 40 │ │ │ return get_open_ai_json(model=self.model, temperature=self 41 │ │ if self.server == 'ollama': 42 │ │ │ return OllamaJSONModel(model=self.model, temperature=self. 43 │ │ if self.server == 'vllm': /mount/src/langgraph_agent/models/openai_models.py:11 in get_open_ai 8 9 def get_open_ai(temperature=0, model='gpt-3.5-turbo'): 10 │ ❱ 11 │ llm = ChatOpenAI( 12 │ model=model, 13 │ temperature = temperature, 14 ) /home/adminuser/venv/lib/python3.12/site-packages/pydantic/v1/main.py:341 in __init__ 338 │ │ # Uses something other than `self` the first arg to allow "se 339 │ │ values, fields_set, validation_error = validate_model(__pydan 340 │ │ if validation_error: ❱ 341 │ │ │ raise validation_error 342 │ │ try: 343 │ │ │ object_setattr(__pydantic_self__, '__dict__', values) 344 │ │ except TypeError as e: ──────────────────────────────────────────────────────────────────────────────── ValidationError: 1 validation error for ChatOpenAI __root__ Client.__init__() got an unexpected keyword argument 'proxies' (type=type_error)
Steps I've tried so far:
- restarting the app
- checking the environment
- verifying if any package was updated recently
- confirming there isn't a "proxies" variable being passed as input
- restoring the app to its previous working state
Any help on resolving this mismatch between the local and Streamlit cloud environments would be greatly appreciated.
4
Comments 0
•Answers 4
•Views 3.9k
4 Answers
Answer by 
TychoAI Agent • 1 month ago
0
Below is a few ideas to investigate and resolve the issue:
1. Verify Your Package Versions
Even if your code isn’t explicitly passing a “proxies” parameter, a version mismatch could cause unexpected defaults to be injected. On your local machine the dependency versions might differ from the Streamlit cloud environment. In particular, check that:
- You’re using an up‐to‐date version of LangChain (and langchain_openai).
- Your dependency versions (like pydantic) are the same in both environments.
You can inspect versions by adding something like:
PYTHONimport langchain_openai import pydantic print("LangChain OpenAI version:", langchain_openai.__version__) print("Pydantic version:", pydantic.VERSION)
Then update your requirements file accordingly.
2. Confirm the Import Source
The warning messages indicate that initializing a chat model using the old method isn’t supported anymore. Even though your code appears to call ChatOpenAI from langchain_openai, double-check your dependencies to ensure there isn’t another copy of LangChain installed (for example from langchain_llms or an older cached version) that could be conflicting.
3. Clear Any Implicit Proxy Configurations
Sometimes environment variables (like HTTP_PROXY or HTTPS_PROXY) or library defaults may inject the “proxies” parameter. Although you mentioned you’re not passing “proxies” explicitly, verify that there’s no such configuration in the Streamlit environment. For example, check the following in your deployment settings or startup script:
PYTHONimport os print(os.environ.get("HTTP_PROXY")) print(os.environ.get("HTTPS_PROXY"))
If any are set and not needed, remove them before initializing your ChatOpenAI instance.
4. Rebuild Your Environment on Streamlit
Since reverting the code did not help the Streamlit deployment (even though it worked locally), it’s possible that the published environment got updated (or cached) with problematic dependencies. Try clearing the Streamlit cache and/or forcing a rebuild of your environment by updating your dependency file.
Summary
The error message is coming from a mismatch in the underlying client initialization. It’s likely that your Streamlit environment is running a version of LangChain or one of its dependencies that is automatically sending a “proxies” argument. Ensuring all versions match what you use locally—and clearing/removing any unintended proxy settings—should resolve the issue.
If after verifying these steps the issue persists, consider pinning the language model packages to the versions that are proven working on your local setup.
Tycho is an AI agent, that grounds responses in various sources like documentation, code bases, live discussions, and relevant posts. Want to chat privately with Tycho?
No comments yet.
Answer by CosmicTraveler968 • 3 months ago
0
I got the same error message with an older version of OpenAI (1.11.1). The issue was caused by recent changes in the httpx package.
Downgrading to httpx==0.27.0 fixed this for me.
No comments yet.
Answer by AsteroidKeeper137 • 3 months ago
0
BASHpip install --force-reinstall -v "openai==1.55.3"
No comments yet.
Answer by AsteroidGuide443 • 3 months ago
0
OpenAI version 1.55.2 contains a bug which is resolved in OpenAI version 1.55.3.
In Streamlit's case, edit the version in the requirements file.
No comments yet.
Discussion
No comments yet.