Logo
Models

Ollama

Learn how to use Ollama to run AI models locally on your computer with CodinIT. Get privacy, no API costs, and offline access to powerful AI.

What is Ollama?

Ollama is a free tool that lets you run AI models directly on your own computer instead of using cloud services. This means your conversations stay completely private, you don't pay API fees, and you can use AI even without internet!Perfect for privacy-conscious developers and those who want full control!

Why Use Ollama with CodinIT?

Complete Privacy

Your code and conversations never leave your computer

No API Costs

Use AI as much as you want without paying per token

Works Offline

Use AI even when you don't have internet connection

Full Control

Choose exactly which models to run and how they behave

What You'll Need

Before setting up Ollama, make sure your computer can handle it:

Good Processor

Modern CPU (Intel i5/AMD Ryzen 5 or better recommended)

Enough RAM

At least 8GB RAM (16GB+ recommended for larger models)

Graphics Card (Optional)

NVIDIA GPU with 4GB+ VRAM for faster performance

Storage Space

5-20GB free space (depending on which models you choose)

Don't worry if your computer isn't super powerful! Ollama has smaller models that work well on most modern computers.

Installing Ollama

Download Ollama

Go to ollama.com and download the installer for your operating system (Windows, Mac, or Linux).

Install and Run

  • Run the installer like any other program
  • Ollama will start automatically in the background
  • You'll see an Ollama icon in your system tray (bottom-right on Windows, top-right on Mac)

Test Your Installation

Open a terminal/command prompt and type ollama --version to make sure it's working.

Choosing Your AI Models

Ollama offers many different models with different capabilities and system requirements:

Beginner-Friendly Models (4-8GB RAM)

ModelSizeIntelligenceBest For
Llama 3.2 3BSmall⭐⭐⭐Quick responses, simple tasks
Phi 3 MiniTiny⭐⭐Very fast, basic assistance
Gemma 2BSmall⭐⭐⭐Balanced performance

Powerful Models (16GB+ RAM)

ModelSizeIntelligenceBest For
Llama 3.1 8BMedium⭐⭐⭐⭐Great all-around performance
CodeLlama 7BMedium⭐⭐⭐⭐Coding and development
Mistral 7BMedium⭐⭐⭐⭐Following instructions precisely

Expert Models (32GB+ RAM)

ModelSizeIntelligenceBest For
Llama 3.1 70BLarge⭐⭐⭐⭐⭐Complex reasoning, best quality
CodeLlama 34BLarge⭐⭐⭐⭐⭐Advanced coding assistance
Start small! Begin with Llama 3.2 3B or Phi 3 Mini to test things out, then upgrade to larger models if you need more capability.

Installing Your First Model

Open Terminal/Command Prompt

  • Windows: Press Win+R, type cmd, press Enter
  • Mac: Press Cmd+Space, type terminal, press Enter
  • Linux: Press Ctrl+Alt+T

Download a Model

Type this command to download a beginner-friendly model:

ollama pull llama3.2:3b

Wait for Download

The model will download (this might take 5-20 minutes depending on your internet speed).

Test Your Model

Once downloaded, test it with:

ollama run llama3.2:3b

Setting Up Ollama in CodinIT

Step 1: Make Sure Ollama is Running

Check that you see the Ollama icon in your system tray and that you've downloaded at least one model.

Step 2: Configure CodinIT

Open Project Settings

In your CodinIT project, click the Settings button (⚙️).

Select Ollama

Choose "Ollama" from the AI Provider dropdown.

Set the Connection

  • Server URL: Usually http://localhost:11434 (this should be the default)
  • Model: Choose the model you downloaded (e.g., llama3.2:3b)

Save Settings

Click Save to apply your changes.

Step 3: Test Your Connection

Try a Simple Prompt

In your CodinIT chat, ask: "Help me create a simple webpage"

Check the Response

If everything works, you'll get a response from your local AI model!

Troubleshoot if Needed

If it doesn't work, make sure Ollama is running and you have the model downloaded.

Managing Your Models

Downloading New Models

# For general use
ollama pull llama3.1:8b

# For coding
ollama pull codellama:7b

# For fast responses  
ollama pull phi3:mini

Storage Management

Check Model Sizes
tip
Use ollama list to see how much space each model uses
Remove Unused Models
tip
Delete models you don't use with ollama rm model-name to free up space
Model Storage Location
tip
Models are stored in your user directory - check Ollama docs for exact location

Understanding Performance

What Affects Speed

Model Size

Smaller models respond faster but may be less capable

Your Hardware

Better CPU/GPU = faster responses

Available RAM

More RAM lets you run larger, smarter models

Performance Tips

  • Use smaller models for quick tasks
  • Close other heavy programs while using AI
  • Consider GPU acceleration if you have NVIDIA graphics
  • Keep conversations shorter for faster responses

Comparing Local vs Cloud AI

When to Use Ollama (Local)

Privacy Matters

Working with sensitive code or personal projects

Cost Control

Heavy usage that would be expensive with API fees

Offline Work

Need to work without internet connection

Full Control

Want to experiment with different models and settings

When to Use Cloud AI

Need Speed

Want the fastest possible responses

Maximum Intelligence

Need the most capable AI models available

Limited Hardware

Computer doesn't have enough power for local models

Easy Setup

Want simple setup without downloading large files

Getting Help

Need more assistance?

Ollama Documentation

Official Ollama guides and model library

Ollama GitHub

Technical documentation and community support

CodinIT Support

Contact our support team for CodinIT-specific help

🏠 Ready to go local? With Ollama connected to your CodinIT project, you now have private, cost-free AI that runs entirely on your computer. Perfect for when privacy and control matter most!

Mistral AI

Learn how to connect Mistral AI models to your CodinIT projects. Get access to powerful AI including specialized coding models like Codestral.

OpenAI

Learn how to connect OpenAI models to your CodinIT projects. Get access to the most popular AI models including GPT-4, o1, and o3.