Run Llama 3 Locally with Ollama: The Complete 2025 Guide

curl -fsSL https://ollama.com/install.sh | sh

ollama run llama3

FROM llama3

# Set the temperature (creativity)
PARAMETER temperature 0.7

# Set the system message
SYSTEM """
You are a Senior React Developer.
You only answer with code snippets and brief explanations.
You prefer Functional Components and Tailwind CSS.
"""

ollama create react-expert -f Modelfile
ollama run react-expert

ollama run dolphin-llama3

const generateResponse = async () => {
  const response = await fetch('http://localhost:11434/api/generate', {
    method: 'POST',
    headers: { 'Content-Type': 'application/json' },
    body: JSON.stringify({
      model: 'llama3',
      prompt: 'Why is the sky blue?',
      stream: false
    })
  });

  const data = await response.json();
  console.log(data.response);
};

import { Ollama } from "@langchain/community/llms/ollama";
import { RetrievalQAChain } from "langchain/chains";

const model = new Ollama({
  baseUrl: "http://localhost:11434",
  model: "llama3",
});

// Imagine 'vectorStore' contains your PDF data
const chain = RetrievalQAChain.fromLLM(model, vectorStore.asRetriever());

const res = await chain.call({
  query: "Summarize the quarterly report based on the PDF."
});

console.log(res.text);