Justme.dev

Appearance

Ollama local AI/LLM integration in IntelliJ Idea and VS-Code

WatchersStarsForksGoView on Github

In today's fast-paced software development environment, integrating advanced AI tools like Ollama can significantly enhance productivity and streamline your workflow. In this blog post, we'll explore how Ollama, a powerful AI assistant integrated into IntelliJ Idea, can transform your coding experience.

What is Ollama? ​

Ollama is an open-source local AI model service that runs on your own computer.

  • Easily download and run Large Language Models
  • Usable on most hardware without dedicated graphics; albeit slower and less accurate
  • Repository with a large selection of specialized models
  • No need to share data with external services
  • Opensource

When installed within IntelliJ Idea, it acts as a bridge between AI models and the code editor.

With Ollama, you can:

  • Access advanced AI tools directly from your IDE
  • Generate high-quality documentation
  • Create well-structured comments with AI-powered suggestions
  • Automate repetitive tasks, like writing commit messages.

Without dedicated hardware it's slightly unbearable running 1.5 billion parameters with deepseek-r1. Simple things like writing small git commit messages work fine, however once the context increases beyond a few 1000 lines of code your really notice the processing limits.

Since it's a network service/api you can just install Ollama on any machine on your local network containing dedicated hardware and host it on your local/vpn network. Check out the Ollama Docker container if you're into that.

Install Ollama: ​

source

bash
curl -fsSL https://ollama.com/install.sh | sh

Start Ollama:

bash
ollama run deepseek-r1:1.5b
#>>> Send a message (/? for help)

To stop output press CTRL+C, type/bye to exit.

Connect your IDE to Ollama ​

IntelliJ Idea ​

Since we have our OpenAI api running with our working language model, lets set up our IDE and start using it. Go to Settings → Tools → AI assistant → Third party AI providers → Enable Ollama.

Set your connection to: http://localhost:11434 (Default) or any other server on your network running Ollama.

Open the IntelliJ AI assistant and set the Model to Ollama - Deepseek-r1:1.5b.

VSCode ​

VSCode users should check out Continue, not tested.

Hardware considerations ​

Depending on your hardware availability, you might want to download bigger models to improve performance and accuracy, check out more Ollama models with more parameters.

Another approach might be to try to with smaller specialized models, switching over to models that fit your specific use case when needed.

Manage models ​

To install new models, e.g. codellama:7b

bash
ollama pull codellama:7b

List installed or remove downloaded models:

bash
ollama list
ollama rm codellama:7b

Uninstall Ollama ​

bash
sudo systemctl disable ollama
sudo rm /etc/systemd/system/ollama.service
sudo rm $(which ollama)
sudo rm -r /usr/share/ollama 
sudo userdel ollama 
sudo groupdel ollama

Conclusion ​

Ollama enhances your workflow in IntelliJ IDEA by providing AI-powered tools without the need for external services or high-end hardware. For deeper insights, consider exploring additional resources from the Ollama's official API documentation.