CodeCraft is a professional AI-powered coding agent that transforms natural language into working software. Unlike basic code generators, CodeCraft maintains conversation memory, executes shell commands, reads and edits existing files, and automatically corrects errors when they occur.
How it works:
Simply describe what you want to build in plain English. CodeCraft writes the code, installs dependencies, runs commands, and delivers a complete production-ready project. Need changes? Just ask. The agent remembers everything and iterates seamlessly.
Live demo: https://codecraft.kawinduwijewardhane.co.uk/
What makes CodeCraft different:

• Persistent conversation memory - the agent remembers what it built
• Self-healing capabilities - when commands fail, CodeCraft reads the error and fixes it
• Real-time visibility - watch every file write, command execution, and fix in the live console
• Complete project automation - from scratch to working application in minutes

Perfect for:

• Developers who want to accelerate their workflow
• Non-technical founders building MVPs
• Freelancers delivering projects faster
• Teams prototyping ideas quickly

Supported technologies:
Next.js, React, Python (Flask/FastAPI), Node.js, TypeScript, Express, HTML/CSS/JS, and any other language or framework.

• Conversational AI Interface - Talk to the agent like a human colleague. Describe what you want to build or change in plain English.
• Persistent Conversation Memory - The agent remembers the entire session. Ask follow-up changes like "change the button color to blue" without re-explaining.
• Real-Time Console Output - Watch every file write, command execution, and error fix as it happens. Full transparency into the agent's work.
• Live File Tree Viewer - See your project structure grow in real-time. Click any file to view its path.
• Automatic Self-Healing - When a command fails, CodeCraft reads the error output and automatically fixes the issue without user intervention.
• Multi-Language Support - Build projects in Next.js, React, Python, Node.js, TypeScript, Express, Flask, FastAPI, static HTML/CSS/JS, and more.
• Full File System Operations - Agent can write, read, edit, list, and delete files within the project directory.
• Shell Command Execution - Run npm install, pip install, python scripts, npx commands, and any terminal command from within the agent.
• Project Management - Create multiple projects, switch between them, and reset conversations. All projects stored in organized folders.
• WebSocket-Based Real-Time Communication - Low-latency communication between browser and server for smooth agent responses.
• Production-Ready Code - No placeholders or partial files. The agent writes complete, working code every time.
• Modern Dark Theme UI - Professional black and green interface designed for developers.

• Node.js 16.0 or higher
• npm 7.0 or higher
• 4GB RAM minimum (8GB recommended)
• 500MB disk space minimum (2GB recommended for projects)
• Windows 10/11, macOS 11+, Ubuntu 20.04+, Debian 11+, or CentOS 8+
• Internet connection for Gemini API calls
• Gemini API key (free from Google AI Studio)
• Chrome, Firefox, Safari, or Edge browser

1. Unzip the downloaded .zip file
2. Follow the instructions in the Documentation folder