LangChain Getting Started

LangChain is a framework for building applications powered by language models. It provides a unified interface across model providers — cloud or local — so the same code that talks to GPT-4 can talk to a locally running Ollama model by swapping one line.

This tutorial walks through installation, environment setup, and the two core invocation patterns: a single blocking call and a streamed token-by-token response.

Prerequisites: Ollama installed and running locally (see the Ollama Setup Guide). Python 3.9+.

Installation and Environment Setup

Installing the Packages

Install all three packages in one command. langchain-ollama provides the ChatOllama integration; python-dotenv handles .env loading.

pip install langchain langchain-ollama python-dotenv

On Linux/macOS: same command — pip or pip3 depending on your Python setup.

python -m venv .venv
.venv\Scripts\activate

Loading Environment Variables

Store your API keys and configuration (LangSmith tracing, endpoints, etc.) in a .env file at the root of your project. Load it at the top of your script:

from dotenv import load_dotenv

load_dotenv('.env')

True

A return value of True confirms at least one variable was loaded from the file. False means the .env file was not found — check that it exists in the working directory.

LANGCHAIN_API_KEY=your_key_here
LANGCHAIN_TRACING_V2=true
LANGCHAIN_PROJECT=my-project

Pulling a Model

Before running any code, make sure the target model is available locally:

ollama pull qwen3

On Linux/macOS: same command — ollama is cross-platform.

Chat with a Local LLM

Creating a ChatOllama Instance

ChatOllama connects to the Ollama server running at localhost:11434. Configure the model name, temperature, and maximum output token count at initialization:

from langchain_ollama import ChatOllama

llm = ChatOllama(
    base_url="http://localhost:11434",
    model="qwen3",
    temperature=0.8,
    num_predict=256
)

Parameter guide:

Invoking the Model

Pass any prompt string to llm.invoke(). It blocks until the full response is ready, then returns a message object — read the generated text from .content:

response = llm.invoke("What is the theory of relativity? Answer in 5 sentences.")
print(response.content)

The theory of relativity, developed by Albert Einstein, consists of two parts: special relativity (1905) and general relativity (1915). Special relativity introduced the idea that the laws of physics are the same for all non-accelerating observers and that the speed of light is constant regardless of the observer's motion. It also established the famous equation E=mc², showing that mass and energy are interchangeable. General relativity extended this to include gravity, describing it not as a force but as the curvature of spacetime caused by mass. Together, they revolutionized our understanding of space, time, and the structure of the universe.

Streaming Responses

For long outputs or real-time UIs, use llm.stream() to receive the response one token chunk at a time instead of waiting for the full generation:

for chunk in llm.stream("What is the theory of relativity? Answer in 5 sentences."):
    print(chunk.content, end="", flush=True)

Inspecting Response Metadata

The response object returned by llm.invoke() exposes runtime statistics via response_metadata. This is useful for debugging latency, tracking token usage, and profiling your prompts:

response.response_metadata

{
  "model": "qwen3",
  "created_at": "2025-10-22T12:16:31.9173234Z",
  "done": true,
  "done_reason": "stop",
  "total_duration": 2630496300,
  "load_duration": 2149414600,
  "prompt_eval_count": 100,
  "prompt_eval_duration": 17446400,
  "eval_count": 140,
  "eval_duration": 397223200,
  "model_name": "qwen3",
  "model_provider": "ollama"
}

Key fields explained:

tps = response.response_metadata['eval_count'] / (response.response_metadata['eval_duration'] / 1e9)
print(f"{tps:.1f} tokens/sec")

Quick Reference

Package Install

pip install langchain langchain-ollama python-dotenv

Minimal Working Example

from dotenv import load_dotenv
from langchain_ollama import ChatOllama

load_dotenv('.env')

llm = ChatOllama(
    model="qwen3",
    temperature=0.8,
    num_predict=256
)

# Single blocking call
response = llm.invoke("Explain transformers in 3 sentences.")
print(response.content)

# Streamed call
for chunk in llm.stream("Explain transformers in 3 sentences."):
    print(chunk.content, end="", flush=True)

ChatOllama Key Parameters