Ollama Integration
Ollama enables fully local AI model execution within OmniIDE. Run coding models on your own hardware with zero cloud dependency and complete privacy.
Setup
- Install Ollama from ollama.ai
- Pull a coding model:
ollama pull codellama:13b - Start the Ollama server:
ollama serve - Configure OmniIDE to use Ollama
Recommended Models
| Model | Size | Best For |
|---|---|---|
| codellama:7b | 3.8 GB | Fast coding on modest hardware |
| codellama:13b | 7.3 GB | Balance of quality and speed |
| deepseek-coder:6.7b | 3.8 GB | Excellent code generation |
Configuration
{
"omni.ollama.enabled": true,
"omni.ollama.endpoint": "http://localhost:11434",
"omni.ollama.model": "codellama:13b"
}Related
- Google Gemini — Cloud AI option
- Ollama Setup Guide
- Offline Mode