← Daily AI recipesStudy✓ Verified Jul 18, 2026

Run an open-source AI model on your computer in 5 minutes

Install Ollama and chat with AI offline without internet or subscriptions

  1. 1.Download the Ollama installer from the official website for your operating system (Windows, macOS, or Linux).

    On the ollama.ai page, click the Download button and download the file for your operating system.

    Ollama ↗
  2. 2.Run the downloaded file and follow the installation instructions. Restart your computer when done.

    On Windows, a standard installer will run. On macOS, drag Ollama to Applications. On Linux, follow the terminal instructions.

    Ollama Installer ↗
  3. 3.Open the command line (cmd on Windows, Terminal on Mac/Linux) and enter the command to download your first model: ollama pull mistral

    Mistral is a fast and compact open model, perfect for your first run. Download takes 2-3 minutes (~4 GB).

    Ollama CLI ↗
    Prompt
    ollama pull mistral
    What this prompt doesThis command downloads the mistral model to your computer. No changes needed - copy and paste as is.
  4. 4.After download, start the local server with this command: ollama serve

    The server starts in the terminal. Keep this window open - it runs in the background.

    Ollama Server ↗
    Prompt
    ollama serve
    What this prompt doesThis command starts the Ollama server at localhost:11434. Copy and paste as is.
  5. 5.Open a new terminal window (without closing the first one) and send a test request to the AI:

    Use the correct /api/chat endpoint to talk with the model. It will return mistral's response.

    curl ↗
    Prompt
    curl http://localhost:11434/api/chat -d '{«model»: «mistral», «messages»: [{«role»: «user», «content»: «[YOUR_QUESTION]»}], «stream»: false}'
    What this prompt doesThis command sends a question to mistral and gets a response. Replace [YOUR_QUESTION] with a real question, like 'How do I cook pasta?' Works on Windows (if curl is installed), Mac, and Linux.
  6. 6.You'll see a JSON response from the AI in the terminal - done! Now use Ollama through a web interface or other applications.

    You can install the Open WebUI separately (docker run -d -p 3000:8080 ghcr.io/open-webui/open-webui:latest) to chat with the model like in ChatGPT.

    Ollama + Open WebUI ↗
💰 What it costs to followprices as of 2026-07
  • freeOllamaCompletely free - no subscription or payments required.
Try it and be the first to check in

Why today

Ollama Officialweb
Open WebUI GitHubweb

AI services change fast - interfaces and free limits may differ from what's described.