Configuration UI Guide
The visual configuration UI provides a web interface for all ChatRAG settings - no code editing required.
Important Note
http://localhost:3000.Accessing the Config UI
Launch the configuration dashboard with:
npm run config
# Opens http://localhost:3333Development Workflow
- Terminal 1:
npm run dev(port 3000) - Main application - Terminal 2:
npm run config(port 3333) - Config dashboard
Configuration Sections
API Keys Tab
Manage all service API keys in one place
Configure:
- Supabase credentials (URL, Anon Key, Service Role Key)
- OpenAI API key for embeddings and models
- OpenRouter API key for 100+ AI models
- LlamaCloud API key for document processing
- FAL.ai, Replicate for media generation
- ElevenLabs for voice features
- Exa for web search
- Stripe/Polar for payments
Features Tab
Enable or disable ChatRAG features
Toggle features:
- Authentication system
- RAG document system
- Image generation (OpenAI, FAL, Replicate)
- Video generation
- 3D model generation
- WhatsApp integration
- MCP tools
- Web search
- Reasoning models
- Embed widget
Branding Tab
Customize your ChatRAG instance appearance
Customize:
- App name and title
- Favicon URL
- Header logo (text or image)
- AI response avatar logo
- Logo sizes and borders
- Custom colors and themes
- Welcome messages and text
- UI preferences
RAG System Prompt
Configure how AI uses your documents
Critical Requirement
{{context}} exactly. Without it, RAG won't work.Features:
- Pre-configured templates (helpful, professional, educational, technical)
- Industry-specific templates (sales, support, research, code assistant)
- WhatsApp-optimized template (concise, conversational)
- AI translation support for multilingual prompts
- Pre-context and post-context configuration
- Real-time preview
Model Management
Add, remove, and configure AI models
Capabilities:
- Add/remove OpenAI models
- Add/remove OpenRouter models
- Fetch latest models from OpenRouter API
- Set default models per feature
- Configure reasoning parameters (effort, tokens)
- Mark models as free or open source
WhatsApp Configuration
Set up WhatsApp integration
Configure:
- Provider selection (Fly.io or Koyeb)
- Provider URL and API key
- Webhook URL and secret
- Enable MCP tools in WhatsApp
- Enable web search in WhatsApp
- Default model for WhatsApp
- Session limits
- Message queue settings
MCP Management
Manage Model Context Protocol integrations
Features:
- Configure Zapier MCP endpoint
- Add custom MCP servers
- Test server connections
- Health monitoring
- YOLO mode toggle (auto-approve tools)
- Enable/disable in chat or embed
- Discover available tools
Admin Tools
User and system management
Access to:
- User management and admin access
- Database inspection tools
- Prompt testing interface
- Document reprocessing
- System health checks
Saving & Applying Changes
After making configuration changes:
- Click "Save Settings" or "Save All Changes" in the Config UI
- Configuration is written to
.env.local - Some settings are also synced to Supabase
admin_settingstable - Restart your development server to apply changes:
# In your dev terminal:
Ctrl+C # Stop the server
npm run dev # Start againConfiguration Persistence
- Changes are saved to
.env.local - Next.js reads environment variables on startup
- Restart required to reload environment changes
- Config UI validates settings before saving
Real-time Testing
Test your configuration changes in real-time:
Make changes in Config UI
Update any settings, API keys, or features
Save configuration
Click save and wait for confirmation
Restart dev server
Reload environment variables
Test in main app
Visit http://localhost:3000 to verify changes
Production Deployment
Config UI is Development Only
For production deployment:
- Configure everything locally using the Config UI
- Copy your finalized
.env.localsettings - Paste environment variables into your hosting provider's dashboard (e.g., Vercel Environment Variables)
- Deploy your application
Pro Tips
- Use the Config UI's "Fetch Models" button to automatically discover new OpenRouter models
- Test your RAG system prompt with the preview feature before saving
- Keep a backup of your
.env.localfile with working settings - Use the validation features to catch configuration errors before deployment