How to Run DeepSeek AI Locally Using Ollama: A Step-by-Step Guide

In our previous guide, we walked you through installing Ollama on Windows, Linux, and macOS to run large language models (LLMs) locally. Now, let’s take the next step: running DeepSeek AI, a cutting-edge model for specialized tasks like coding assistance, data analysis, or creative writing. In this tutorial, you’ll learn how to use Ollama to deploy DeepSeek AI on your machine and start leveraging its capabilities in minutes. Additionally, I will show you how to run DeepSeek AI locally, ensuring you can maximize its potential.

Prerequisites

Before proceeding, ensure you’ve completed these steps:

1. Install Ollama: Follow our detailed guide to set up Ollama on your OS (Windows, Linux, or macOS).
2. System Requirements: At least 8GB RAM (16GB recommended) and 20GB of free storage for larger models.

Step 1: Pull the DeepSeek AI Model

Ollama simplifies model management with its command-line interface. Deepseek has different model but in this guide we will be using deepseek-coder-v2. This is trained specially for coding task. To download DeepSeek AI model, open your terminal (or PowerShell/Command Prompt on Windows) and run:

ollama pull deepseek-coder-v2

This command fetches the latest version of deepseek-coder-v2 from Ollama’s model library.

Note: If the model isn’t found, check the Ollama model registry for the exact name or alternative versions.

Step 2: Run DeepSeek AI Locally

Once you have everything set up, you’re ready to run DeepSeek AI locally and explore its features.

Once the model is downloaded, start interacting with it using:

ollama run deepseek-coder-v2

This launches an interactive chat session. For example, you can ask:

>>> Write a Python function to calculate Fibonacci numbers.  

DeepSeek AI will generate code or answers in real-time.

Step 3: Customize Your Experience

Save Conversations:

Redirect output to a file:

ollama run deepseek-coder-v2 "Explain quantum computing" > output.txt  

Troubleshooting

• Model Not Responding? Ensure Ollama’s service is running:
  • Windows/Linux: ollama serve (in the background).
  • macOS: The app runs automatically after installation.
• Out of Memory? Close other resource-heavy apps or try smaller model variants (e.g., deepseek-r1-7b instead of deepseek-r1-70b).

Remember, running DeepSeek AI locally allows for greater flexibility and control over your AI projects.

By learning to run DeepSeek AI locally, you can integrate powerful AI into your projects with ease.

Conclusion

Running DeepSeek AI locally with Ollama unlocks powerful AI capabilities right on your Windows, Linux, or macOS machine. By following this guide, you’ve learned to install, customize, and integrate the model for coding, research, or creative projects. For more tips, revisit our guide on how to install Ollama, and explore Ollama’s model library to experiment with other LLMs!

Have questions? Let us know how you run DeepSeek AI locally and share your experiences!

FAQs