What it is
Ollama is a free, open-source program that downloads open-weight language models (Llama, Gemma, Qwen, Mistral, DeepSeek and many others) and runs them on your own machine. Nothing you type leaves your computer, it works offline once a model is downloaded, and there is no usage bill. It also starts a local API so other apps (Open WebUI, VS Code extensions, your own scripts) can use the models.
Install
Windows and macOS: download the installer from ollama.com/download and run it. Ollama then runs in the background (look for its icon in the menu bar / system tray).
Linux: run the official install script:
curl -fsSL https://ollama.com/install.sh | shCheck it works:
ollama --versionUse it
Download and chat with a model in one command (the first run downloads it):
ollama run llama3.2Type your message, press Enter, and type /bye to quit. Other useful commands:
| Command | What it does |
|---|---|
ollama pull gemma3 | Download a model without starting a chat |
ollama list | Show the models on your disk and their size |
ollama ps | Show which models are loaded in memory |
ollama rm gemma3 | Delete a model to free disk space |
ollama run qwen3 "Summarise this: ..." | One-off question without opening a chat |
Browse all available models (with sizes) in the Ollama library. Model names change often, so check there for the latest versions.
Tip
Start with a small model (1–4 billion parameters). If answers are fast, try a bigger one. See Choose the right local model for your computer.
Use the local API
Ollama listens on http://localhost:11434. Any program can send it a request:
curl http://localhost:11434/api/chat -d '{
"model": "llama3.2",
"messages": [{ "role": "user", "content": "Give me 3 names for a bakery" }],
"stream": false
}'It also offers an OpenAI-compatible endpoint at http://localhost:11434/v1, so many tools built for ChatGPT work by just changing the address.
Limitations
- Local models are smaller than the best cloud models: great for drafting, summarising and coding help, weaker on hard reasoning.
- Speed depends on your hardware. A recent GPU or an Apple Silicon Mac makes a big difference.
- Models take several gigabytes of disk space each.
Step by step: Install Ollama and run your first local AI.
Guides for Ollama
Install Ollama and run your first local AI
Run a ChatGPT-style AI entirely on your own computer, for free and offline, in about 15 minutes.
Uses: Ollama
Choose the right local model for your computer
Understand parameters, quantization and memory, and find out which local AI models your computer can run comfortably.
Uses: Ollama, LM Studio, llama.cpp
Get a ChatGPT-like interface for local AI (Open WebUI + Ollama)
Install Open WebUI on top of Ollama to get your own private ChatGPT in the browser, with history, document chat and multiple users.
Uses: Open WebUI, Ollama
Comments 0
Share tips, prompts that worked for you, or questions.
No comments yet. Be the first to share a tip.
Log in or create a free account to comment.