Ollama is a tool that lets you run large language models (LLMs) locally on your machine instead of calling cloud APIs.

What it actually does

• Downloads AI models (like LLaMA, Mistral, DeepSeek)
• Runs them on your CPU/GPU
• Exposes a simple CLI + REST API
• No internet needed after models are pulled

Why people use it

• Privacy → everything stays local
• No API costs
• Fast iteration for dev work
• Works well for:
  • code generation
  • chatbots
  • local automation
  • embedding + RAG systems

Basic usage (what you care about)

Pull a specific model

Call it from your app (like your Bootstrap modal app)

Models you can run

Common ones:

• llama3 (general)
• mistral (fast)
• deepseek-coder (code)
• qwen2.5-coder (strong coding)
• phi (lightweight)

Important limitation

• Runs smaller/open models, not things like GPT-5
• Quality depends on:
  • your hardware (RAM/VRAM)
  • model size (7B, 13B, 70B, etc.)

Bottom line

Ollama = local AI engine
Think:

“Docker for LLMs running on my machine”