A lightweight, privacy-focused interface for unlimited AI conversations

As someone who frequently uses Claude AI, I often encountered the conversation limit issue that disrupted my workflow. I created LLMBox to solve this problem - allowing for longer, uninterrupted conversations without hitting Claude's usual limits. While the current version works exclusively with Claude's API, I designed it to maintain continuous dialogue by efficiently managing conversation context.

Note: I'm not a professional developer - I'm a growth enthusiast who loves building solutions to real problems. While LLMBox might not be perfect from a coding perspective, it effectively serves its primary purpose. I welcome suggestions and contributions from experienced developers to improve the codebase!

• 🎨 Modern, clean UI with dark/light mode support
• 📁 File upload and analysis capabilities
• 🔍 Full-text search in conversations and messages
• 💾 Local storage for chat history
• ⚡️ Fast and responsive design
• 📱 Mobile-friendly interface
• 🛡️ Built-in rate limiting and error handling
• 🔄 Real-time message streaming
• 🎛️ Advanced settings control (temperature, models)

graph TD
    A[React Frontend] -->|API Requests| B[Node.js Backend]
    B -->|API Calls| C[Claude API]
    B -->|File Storage| D[Local Storage]
    A -->|State Management| E[Zustand Store]
    B -->|File Processing| F[Multer]

1. Unlimited Conversations

   • Smart context management
   • Continuous dialogue flow
   • No artificial limits

2. Privacy First

   • Local data storage
   • No cloud dependency
   • Complete data control

3. Enhanced Features

   • File analysis
   • Full-text search
   • Custom AI parameters
   • Real-time streaming

• React.js for UI
• Tailwind CSS for styling
• Zustand for state management
• Lucide Icons for UI elements

• Node.js runtime
• Express.js framework
• Multer for file handling
• Claude API integration

• Node.js (v14 or higher)
• npm or yarn
• Claude API access
• Modern web browser

1. Clone and Setup

git clone https://github.com/itsbariscan/llmbox.git
cd llmbox

2. Install Dependencies

# Frontend
cd frontend && npm install

# Backend
cd ../backend && npm install

3. Configure Environment

cp .env.example .env
# Add your Anthropic API key to .env

4. Launch

# Backend
npm start

# Frontend
cd ../frontend && npm start

Error: Invalid API key
Solution: Ensure ANTHROPIC_API_KEY is correctly set in .env

Error: ECONNREFUSED
Solution: Check if backend server is running on correct port

Error: File too large
Solution: Adjust MAX_FILE_SIZE in environment variables

1. Optimal Context Management

   • Keep context length under 4000 tokens
   • Clear unnecessary conversation history
   • Use file analysis sparingly

2. Browser Performance

   • Enable hardware acceleration
   • Clear browser cache regularly
   • Use latest Chrome/Firefox/Safari

I'm planning to add:

• Multi-provider LLM support
• User authentication system
• Conversation export/import
• Collaborative features
• Voice interaction
• Custom theming
• Improved error handling
• Better code organization

# Run tests
npm test

# Build production
npm run build

# Run linter
npm run lint

# Format code
npm run format

As a growth enthusiast rather than a professional developer, I especially welcome contributions that can improve the code quality and add new features. Whether you're a seasoned developer or a fellow enthusiast, your input is valuable!

• Code optimization
• Feature development
• Documentation
• Bug fixes
• UI/UX improvements

• Claude API exclusivity
• Basic error handling
• Limited security features
• No test coverage
• Basic rate limiting

Hi, I'm Baris Can, an SEO & Organic Growth Enthusiast focused on growth marketing strategies, automation, and coding. My expertise lies in optimizing digital presence and driving growth through data-driven approaches.

• SEO and Growth Marketing
• Marketing Automation
• Python and JavaScript programming
• Blockchain technologies
• Building practical solutions

• GitHub: @itsbariscan
• LinkedIn: linkedin.com/in/bariscansayin
• Twitter: @bcsayin
• Email: seo.bariscan@gmail.com
• SubStack: https://itsbariscan.substack.com/

I'm grateful to:

• Anthropic for Claude API
• Open source community

This project is born from enthusiasm rather than professional software development expertise. While functional, there's always room for improvement. Feedback and contributions are welcome!