Run Large Language Models (LLMs) Locally

Tired of paying for API access just to experiment with LLMs? You can run powerful open-source models directly on your own machine — no external dependencies, no cost, and your data never leaves your computer.

Why Run LLMs Locally?

• Privacy — your prompts and data stay on your machine
• Cost — completely free, no API billing
• Development — build and test AI-powered applications without rate limits or latency from remote servers

Method 1: Ollama

Ollama is a CLI tool that makes downloading and running language models as simple as a single command.

Install Ollama from ollama.com/download, then run any model:

ollama run deepseek-r1

That's it. Ollama pulls the model on first run and starts an interactive chat session in your terminal. You can swap deepseek-r1 for any model from the Ollama library:

# Run Llama 3
ollama run llama3
 
# Run Mistral
ollama run mistral
 
# Run Phi-3
ollama run phi3

Ollama also exposes a local REST API on http://localhost:11434, so you can integrate it directly into your applications:

curl http://localhost:11434/api/generate -d '{
  "model": "deepseek-r1",
  "prompt": "Explain how transformers work in simple terms",
  "stream": false
}'

Or use it from Node.js:

const response = await fetch("http://localhost:11434/api/generate", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "deepseek-r1",
    prompt: "Explain how transformers work in simple terms",
    stream: false,
  }),
});
 
const data = await response.json();
console.log(data.response);

Method 2: LM Studio

LM Studio offers a graphical interface — think of it as a local ChatGPT running entirely on your machine.

1. Download from lmstudio.ai

2. Open the Discover tab and search for a model

3. Download your chosen model
4. Switch to the Chat tab, select the model, and start chatting

LM Studio also includes a local server mode that emulates the OpenAI API, so you can drop it into any app already using the OpenAI SDK with zero code changes.

Ollama vs LM Studio

Conclusion

Running LLMs locally is easier than ever. Whether you prefer the simplicity of a terminal command with Ollama or the ChatGPT-like experience of LM Studio, you can have a powerful AI model running on your own hardware in minutes — for free.