Ollama as the Model Layer in Your Agent Stack: LangChain, LlamaIndex, and n8n Integration

from langchain_ollama import ChatOllama, OllamaEmbeddings
 
# Chat model
llm = ChatOllama(
    model='llama3.1:8b',
    base_url='http://localhost:11434',  # default; can omit
    temperature=0,
)
 
response = llm.invoke('Explain what a vector database is in one paragraph.')
print(response.content)
 
# Embeddings (for RAG)
embeddings = OllamaEmbeddings(model='nomic-embed-text')
vector = embeddings.embed_query('What is a transformer model?')
print(f'Embedding dimension: {len(vector)}')  # e.g. 768
 

# Full RAG chain with Ollama
from langchain_ollama import ChatOllama, OllamaEmbeddings
from langchain_community.vectorstores import Chroma
from langchain.chains import RetrievalQA
from langchain_community.document_loaders import DirectoryLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
 
# Load and chunk documents
loader = DirectoryLoader('./docs/', glob='**/*.md')
docs = loader.load()
splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=50)
chunks = splitter.split_documents(docs)
 
# Embed and store locally
embeddings = OllamaEmbeddings(model='nomic-embed-text')
vectorstore = Chroma.from_documents(chunks, embeddings, persist_directory='./chroma_db')
 
# RAG chain
llm = ChatOllama(model='llama3.1:8b', temperature=0)
chain = RetrievalQA.from_chain_type(
    llm=llm,
    retriever=vectorstore.as_retriever(search_kwargs={'k': 4}),
)
 
answer = chain.invoke('What is the refund policy?')
print(answer['result'])
 

from llama_index.llms.ollama import Ollama
from llama_index.embeddings.ollama import OllamaEmbedding
from llama_index.core import VectorStoreIndex, SimpleDirectoryReader, Settings
 
# Configure global settings
Settings.llm = Ollama(model='llama3.1:8b', request_timeout=120.0)
Settings.embed_model = OllamaEmbedding(model_name='nomic-embed-text')
 
# Load documents and build index
documents = SimpleDirectoryReader('./docs').load_data()
index = VectorStoreIndex.from_documents(documents)
 
# Query
query_engine = index.as_query_engine()
response = query_engine.query('Summarise the key product features.')
print(response)
 

Set request_timeout=120.0 or higher for Ollama with LlamaIndex. Larger local models can take 30-60 seconds to respond on the first call (cold start) or for long documents. The default timeout is too short and causes confusing errors.

{
  "baseUrl": "http://host.docker.internal:11434",
  "model": "nomic-embed-text"
}
 

If n8n and Ollama are both in Docker Compose (not recommended — Ollama should run directly on the GPU host), use the service name as the host: http://ollama:11434. The host.docker.internal pattern only works when Ollama runs on the Docker host machine.

from openai import OpenAI
 
client = OpenAI(
    base_url='http://localhost:11434/v1',
    api_key='ollama',  # required by the client; value ignored by Ollama
)
 
response = client.chat.completions.create(
    model='qwen2.5-coder:7b',
    messages=[{'role': 'user', 'content': 'Write a Python function to reverse a string'}],
)
print(response.choices[0].message.content)