TzamunCode CLI 🚀

AI Coding Assistant powered by local models - Built in Saudi Arabia 🇸🇦

TzamunCode is a privacy-first AI coding assistant that runs entirely on your local infrastructure using Ollama and vLLM. No cloud dependencies, no API costs, complete control.

✨ Features

• 🤖 Agentic AI - Multi-step planning and execution
• 📝 Code Generation - Create files, functions, and entire projects
• ✏️ Multi-file Editing - Edit multiple files in one operation
• 🔧 Tool Calling - Git operations, file search, command execution
• 🎯 Context Aware - Understands your project structure
• 🔒 Privacy First - Everything runs locally
• ⚡ Fast - Powered by vLLM for optimized inference
• 🌍 Multi-model - Use any Ollama model (15+ available)

🚀 Quick Start

Installation

# Clone the repository
git clone https://github.com/tzamun/tzamuncode-cli.git
cd tzamuncode-cli

# Install
pip install -e .

# Or install from PyPI (when published)
pip install tzamuncode

Prerequisites

• Python 3.9+
• Ollama running locally (http://localhost:11434)
• At least one Ollama model installed

Basic Usage

# Start interactive chat
tzamuncode chat

# Generate code
tzamuncode generate "Create a Flask REST API with authentication"

# Edit a file
tzamuncode edit app.py "Add error handling to all routes"

# Explain code
tzamuncode explain main.py

# Quick alias
tzc chat

📖 Documentation

Commands

chat - Interactive Chat

tzamuncode chat
tzamuncode chat --model qwen2.5:32b

generate - Code Generation

tzamuncode generate "Create a Python web scraper"
tzamuncode generate "Add unit tests for user.py" --output tests/

edit - File Editing

tzamuncode edit app.py "Refactor to use async/await"
tzamuncode edit . "Add type hints to all functions"

explain - Code Explanation

tzamuncode explain complex_function.py
tzamuncode explain --detailed auth.py

review - Code Review

tzamuncode review pull_request.diff
tzamuncode review --strict src/

Configuration

Create ~/.tzamuncode/config.yaml:

# Default model
model: qwen2.5:32b

# Ollama settings
ollama:
  base_url: http://localhost:11434
  timeout: 120

# vLLM settings (optional, for faster inference)
vllm:
  enabled: true
  base_url: http://localhost:8000
  model: deepseek-coder-7b

# Preferences
preferences:
  show_diff: true
  auto_apply: false
  max_context: 64000
  temperature: 0.7

🏗️ Architecture

┌─────────────────────────────────────┐
│     TzamunCode CLI                  │
│  (Typer + Rich UI)                  │
└─────────────────────────────────────┘
           ↓
┌─────────────────────────────────────┐
│   Agentic Layer (LangChain)         │
│  - Multi-step planning              │
│  - Tool calling                     │
│  - Context management               │
└─────────────────────────────────────┘
           ↓
┌─────────────────────────────────────┐
│   AI Backend                        │
│  - Ollama (15+ models)              │
│  - vLLM (fast inference)            │
└─────────────────────────────────────┘

🤝 Contributing

We welcome contributions! See CONTRIBUTING.md for guidelines.

📄 License

MIT License - see LICENSE for details.

🌟 Built by Tzamun Arabia IT Co.

TzamunCode is part of the Tzamun AI ecosystem:

• TzamunAI - AI platform with 15+ models
• TzamunERP - ERPNext + AI integration
• Auxly - AI coding assistant for IDEs
• AccessHub - Privileged Access Management

Visit tzamun.com to learn more.

Made with ❤️ in Saudi Arabia 🇸🇦