infrastructure-ollama-configuration-chatollama-from-langchain-ollama-does-not-accept-a-b7afe8ab

THEN Set the OLLAMA_HOST environment variable to the Ollama server URL before importing langchain_ollama. If the server is behind a proxy (e.g., Open Web UI), ensure the variable points to the root Ollama endpoint (e.g., http://host:port) without a subpath. Alternatively, use langchain_community.llms.ollama.Ollama which supports a base_url parameter.

import os
os.environ["OLLAMA_HOST"] = "http://my-ollama-server:11434"
from langchain_ollama import ChatOllama
llm = ChatOllama(model="llama3")