An autonomous agentic coding assistant powered by Ollama with native tool calling. Similar to OpenCode or Claude Code, but using local LLMs through Ollama. Works completely offline and respects your privacy.

• 🤖 Autonomous Agentic Coding - Works independently to complete tasks without constant supervision
• 🛠️ Native Ollama Tool Calling - Uses Ollama's built-in tool calling capabilities for structured function execution
• 📁 File Operations - Read, write, create, delete files and directories
• 💻 Terminal Commands - Execute shell commands (cd, ls, git, npm, pip, etc.)
• 🌐 Web Search - Search the internet using DuckDuckGo (no API keys required)
• 🌍 Browser Automation - Control Chrome/Selenium for web interactions and scraping
• 💬 Beautiful Terminal UI - Rich library-based interface with syntax highlighting
• 🔒 Privacy-First - All processing happens locally, no data sent to external services
• 🚀 Fast & Efficient - Uses local LLM inference, works offline

• Python 3.8+ (python3)
• Ollama - Install from ollama.ai
• Chrome/Chromium and ChromeDriver (optional, for browser automation)
  • Arch Linux: sudo pacman -S chromium chromium-driver
  • Ubuntu/Debian: sudo apt-get install chromium-browser chromium-chromedriver
  • macOS: brew install chromium chromedriver

git clone https://github.com/r3dg0d/ollamacode.git
cd ollamacode
chmod +x setup.sh
./setup.sh

The setup script will:

• Create a Python virtual environment
• Install all dependencies
• Check for Ollama installation
• Optionally pull the default model
• Create a launcher script

git clone https://github.com/r3dg0d/ollamacode.git
cd ollamacode

# Create virtual environment
python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install --upgrade pip
pip install -r requirements.txt

# Make launcher executable
chmod +x ollamacode

# Using yay
yay -S ollamacode

# Using paru
paru -S ollamacode

# Or manually
git clone https://aur.archlinux.org/ollamacode.git
cd ollamacode
makepkg -si

Add to your PATH:

echo 'export PATH="$PATH:$HOME/Documents/ollamacode"' >> ~/.bashrc
source ~/.bashrc

Or create a symlink:

ln -s ~/Documents/ollamacode/ollamacode ~/.local/bin/ollamacode

# Start Ollama (if not already running)
ollama serve

# Pull the recommended model (if not already installed)
ollama pull huihui_ai/qwen3-abliterated:32b

Note: You can use any Ollama model that supports tool calling. Check Ollama's model library for compatible models.

Interactive Mode:

ollamacode
# or
python main.py

Autonomous Mode:

ollamacode --task "Create a Python script that prints 'Hello World'"
# or
python main.py --task "Create a new React project with TypeScript"

ollamacode [OPTIONS]

Options:
  --model MODEL          Ollama model to use 
                         (default: huihui_ai/qwen3-abliterated:32b)
  
  --endpoint URL         Ollama API endpoint 
                         (default: http://127.0.0.1:11434)
  
  --base-dir PATH        Base directory for file operations 
                         (default: ~/Documents)
  
  --task TASK            Initial task to execute (autonomous mode)

ollamacode --task "Create a new Python project called 'myapp' with a main.py file and a requirements.txt"

ollamacode --task "Search for information about Python decorators and create a file with examples"

ollamacode --task "Open a browser, navigate to example.com, and save the page HTML to a file"

ollamacode
# Then type your tasks interactively
Task: Create a Flask REST API with endpoints for users
Task: Add authentication middleware
Task: Write unit tests

The assistant has access to these tools:

• read_file(filepath) - Read file contents
• write_file(filepath, content) - Write to a file (creates parent directories)
• create_directory(dirpath) - Create directories
• list_directory(dirpath) - List directory contents
• delete_file(filepath) - Delete a file
• delete_directory(dirpath, recursive) - Delete directories

• run_command(command, cwd) - Execute shell commands (cd, ls, git, npm, pip, etc.)

• web_search(query, max_results) - Search the web using DuckDuckGo
• open_browser(headless) - Open Chrome browser
• navigate_to(url) - Navigate to URL
• browser_click(selector, by) - Click element
• browser_type(selector, text, by) - Type text
• browser_get_text(selector, by) - Get element text
• browser_get_html() - Get page HTML
• close_browser() - Close browser

1. You provide a task or question - Either interactively or via command-line
2. The assistant uses Ollama with tool calling - Plans and executes actions autonomously
3. Tools are executed automatically - File operations, commands, web searches, etc.
4. Results are fed back to the model - The model processes results and continues
5. The process continues until the task is complete - Up to 50 iterations per task

Edit main.py or pass via command-line:

ollamacode --model llama3.2:3b --task "your task"

ollamacode --base-dir ~/Projects --task "create a new project"

If Ollama is running on a different host/port:

ollamacode --endpoint http://192.168.1.100:11434 --task "your task"

Problem: Failed to connect to Ollama at http://127.0.0.1:11434

Solution:

# Make sure Ollama is running
ollama serve

# Check if Ollama is accessible
curl http://127.0.0.1:11434/api/tags

Problem: Selenium/Chrome errors when using browser tools

Solution:

# Install Chrome/Chromium and ChromeDriver
# Arch Linux
sudo pacman -S chromium chromium-driver

# Ubuntu/Debian
sudo apt-get install chromium-browser chromium-chromedriver

# macOS
brew install chromium chromedriver

Problem: Error: model not found

Solution:

# Pull the model
ollama pull huihui_ai/qwen3-abliterated:32b

# Or use a different model
ollamacode --model llama3.2:3b --task "your task"

Problem: Permission denied when accessing files/directories

Solution:

• Check file/directory permissions in the base directory
• Ensure the base directory path is correct
• Use absolute paths if relative paths fail

Problem: Module not found errors

Solution:

# Make sure virtual environment is activated
source venv/bin/activate  # Linux/macOS
# or
venv\Scripts\activate  # Windows

# Reinstall dependencies
pip install -r requirements.txt

ollamacode/
├── main.py              # Main entry point
├── ollama_client.py     # Ollama API client with tool calling
├── tools.py             # Tool implementations (file ops, terminal, web)
├── ui.py                # Terminal UI using Rich library
├── ollamacode           # Launcher script (bash)
├── setup.sh             # Setup script
├── requirements.txt     # Python dependencies
├── README.md            # This file
├── PKGBUILD             # AUR package build file
└── LICENSE              # MIT License

# Activate virtual environment
source venv/bin/activate

# Run directly with Python
python main.py

# Or use the launcher
./ollamacode

1. Add tool implementation to tools.py in the ToolRegistry class
2. Add tool definition to get_tool_definitions() method
3. Tools are automatically available to the LLM

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

OllamaCode works with any Ollama model that supports tool calling. Recommended models:

• huihui_ai/qwen3-abliterated:32b - Default, excellent coding capabilities
• llama3.2:3b - Lightweight, fast
• codellama:7b - Code-specialized model
• deepseek-coder:6.7b - Strong coding model

Check Ollama's model library for more options.

• 100% Local - All processing happens on your machine
• No Data Collection - No telemetry, no analytics, no external requests (except web search)
• No API Keys Required - Uses local Ollama instance
• File System Access - Be careful with commands, the assistant can read/write files

⚠️ Warning: This tool can execute arbitrary shell commands and modify your file system. Only use with trusted models and tasks.

MIT License - see LICENSE file for details

• Ollama - Amazing local LLM runtime
• Rich - Beautiful terminal UI
• Selenium - Browser automation
• DuckDuckGo Search - Web search

• Issues: GitHub Issues
• Discussions: GitHub Discussions

Made with ❤️ by r3dg0d