How to Install DeepSeek R1 Locally

How to Install DeepSeek R1 Locally

January 30, 2024•ArticFlow Team

AI is evolving fast, and so are the tools that power it. OpenAI's models are great, but they come at a cost—both financially and in terms of privacy. Enter DeepSeek R1, a reasoning model that rivals OpenAI's GPT-4, but runs completely on your local machine. No cloud. No subscriptions. No data leaks.

Let's dive into the installation process. Don't worry if you're new to this - we'll go through each step carefully.

Step 1: Install Olama (The Core Engine)

DeepSeek R1 runs on a lightweight framework called Olama, which we need to install first.

  1. Go to olama.com
  2. Download Olama for your OS (Windows 10+, macOS, or Linux)
  3. Install and unzip the file
  4. Run the installer (just follow the prompts)

Once installed, you'll see a command line window pop up. That's your Olama terminal—where all the magic happens.

Step 2: Install DeepSeek R1 Model

With Olama installed, it's time to load up DeepSeek R1.

  1. Go to the Models tab on olama.com
  2. Select DeepSeek R1
  3. Copy the installation command (it will look like olama run deepseek-r1)
  4. Open the terminal and paste the command
  5. Press Enter

Tip:

Start with the 7B model (4.7GB). If you have a high-end GPU, you can try larger models, but for most users, 7B is the sweet spot.

Step 3: (Optional) Install Llama 3.3

Want another open-source model for comparison? Install Llama 3.3 by running:

olama run llama3-3.3

This gives you a general-purpose LLM alongside DeepSeek's reasoning model.

Step 4: Install Docker (For the UI)

Using the command line is great, but let's make it user-friendly. We'll install Docker, which allows us to run a ChatGPT-style web UI.

  1. Go to docker.com
  2. Download & install Docker for your OS
  3. Run Docker and keep it open in the background

That's it! No need to configure anything.

Step 5: Install Open WebUI

Now, let's install a beautiful chat interface so you don't have to use the command line.

  1. Go to docs.openwebui.com
  2. Copy the installation command (it will look like olama openwebui)
  3. Open terminal and paste the command
  4. Press Enter

When done, open Docker → find the Open WebUI container → copy the port number (e.g., localhost:3000). Paste it into your browser.

You now have a private ChatGPT alternative running locally!

Final Thoughts

With DeepSeek R1 installed, you now have a powerful AI assistant that doesn't rely on the cloud. No API calls, no subscription fees—just pure, local AI.

I'll be comparing DeepSeek R1 to OpenAI's models in an upcoming post, but in the meantime, try it out and let me know your thoughts!

←Back to all articles