A hybrid Electron + React desktop application where users select and organize their own context for AI-assisted chat. It allows users to highlight text from files, organize it into context folders, and chat with an AI model that answers based solely on the selected context.
- Multi-tab interface: File Manager and Chat
- Highlight text in files and save snippets as context
- Organize snippets into named folders
- Chat with AI using only selected context
- Add, edit, and delete context folders
- Supports .txt and .docx files (basic support for .doc)
- Electron desktop application
- View chat log
upload, highlight text, and create context groups
chat with user selected context
Prerequisites:
- Node.js >= 20
- npm or yarn
- Python >= 3.10 (for backend API)
- api token from HuggingFace
npm installpip install flask flask-cors python-dotenv huggingface-hubRun Development Server:
- Start backend:
python app.py- Start Vite server:
npm run dev- Start Electron app: (must be running at the same time as vite server)
npm run electronFile Manager Tab
- Add .txt or .docx files
- Highlight text to create context snippets
- Save snippets into named context folders
- Edit or delete folders
Chat Tab
- Select a context folder
- Ask questions based only on the selected context
- Start a new chat session anytime
- React
- Vite
- Electron
- SCSS
- Python
- Flask
- HuggingFace Inference API


