opencode-ollama

AI coding agent with native Ollama support. Fork of OpenCode modified to work seamlessly with any Ollama server.

Features

• 🚀 Native Ollama integration - no hacks or workarounds
• 🌐 Connect to any Ollama master server
• 🤖 Auto-detects available models on your Ollama server
• 💻 Works with local or remote Ollama instances
• ⚡ Same great OpenCode TUI experience

Installation

npm install -g opencode-ollama

Usage

Connect to Local Ollama

opencode-ollama

Connect to Remote Ollama Server

OLLAMA_HOST=http://192.168.1.100:11434 opencode-ollama

Set Default Server

Add to your .bashrc or .zshrc:

export OLLAMA_HOST=http://your-ollama-server:11434

Then just run:

opencode-ollama

Supported Models

Works with any model available on your Ollama server:

• qwen3:32b, qwen3.5:35b
• deepseek-r1
• llama3, codellama
• mistral, mixtral
• Any custom fine-tuned models

Ollama Setup

Make sure Ollama is configured to accept network connections:

# On your Ollama server
OLLAMA_HOST=0.0.0.0:11434 ollama serve

Or via systemd:

sudo systemctl edit ollama
# Add: Environment="OLLAMA_HOST=0.0.0.0:11434"
sudo systemctl restart ollama

Original Project

This is a fork of OpenCode with added Ollama support.

License

MIT