A powerful cross-platform Electron desktop application that combines local AI inference with an intuitive code editor. Built with QVAC SDK for offline AI capabilities, React + Vite for the UI, and an embedded Express server for script management. Generate, edit, and run scripts with the help of a locally-running AI assistant.
-
Node.js 20+ and npm
Install Node.js | Node.js Setup Guide -
Optional Runtimes: Install languages you plan to use:
- Python 3.8+ for
.py
scripts
Install Python | Python Setup Guide - Node.js for
.js
/.ts
scripts
Install Node.js | Node.js Setup Guide - Bash/Zsh for shell scripts
Bash Installation | Zsh Installation | Shell Setup Guide - Go, Rust, Java, etc. for respective file types
Install Go | Go Setup Guide
Install Rust | Rust Setup Guide
Install Java | Java Setup Guide
- Python 3.8+ for
git clone https://github.com/your-username/coding-ai-agent.git
cd coding-ai-agent
npm run setup
Start the React dev server and Electron together:
npm run dev
This runs:
renderer
: Vite dev server athttp://localhost:5173
electron
: Electron with live reload, waiting for the renderer
- Slow Loading: First-time model download can take 5-10 minutes
- Memory Errors: Ensure at least 8GB RAM available
- Model Not Found: Delete
~/.qvac
folder and restart app
- Permission Denied: Check file permissions in storage directory
- Command Not Found: Install missing dependencies (Python, Node.js, etc.)
- Path Issues: Ensure storage directory is accessible