MinutesMaster AI is a sophisticated web application for generating professional meeting minutes using AI technology. Transform meeting transcripts into structured, professional documentation with customizable terminology, organizational style preferences, and multi-provider AI support.
- AI Pipeline: Complete implementation supporting OpenRouter, OpenAI, Anthropic, and custom endpoints
- Smart Context Building: Combines user dictionaries, instructions, and sample minutes for consistent output
- Structured Minutes Generation: Professional format with attendees, decisions, action items, and next steps
- Real-time Progress: Live status updates and progress tracking during generation
- Setup Wizard: Guided onboarding for transcript input, sample upload, and configuration
- Three-Panel Workspace: Professional layout with transcript, minutes, and customization tools
- Dark/Light Theme: Toggle between themes with system preference detection
- Meeting History: Track and access previous generated minutes
- OpenRouter: Access to multiple AI models through unified API
- OpenAI: Direct integration with GPT models
- Anthropic: Claude model support via compatible endpoints
- Custom Endpoints: Support for any OpenAI-compatible API
- Node.js 16+
- Modern browser with ES2020+ support
- AI API key from supported provider (OpenRouter, OpenAI, etc.)
# Install dependencies
npm install
# Start development server
npm run dev
# Open browser to http://localhost:5173# Build application
npm run build
# Preview production build
npm run preview# Run tests (when test framework is configured)
npm test
# Run linting
npm run lint- Click the settings gear icon in the workspace
- Select your preferred AI provider:
- OpenRouter: Recommended for access to multiple models
- OpenAI: Direct OpenAI API access
- Custom: Any OpenAI-compatible endpoint
- Enter your API key
- Test the connection
- Adjust model parameters as needed
For professional meeting minutes:
- Temperature: 0.4 (focused, consistent output)
- Top P: 0.9 (balanced diversity)
- Max Tokens: 4000 (adequate for detailed minutes)
The core AI service handles:
- Multi-provider API management
- Context building from user data
- Structured prompt engineering
- Response validation and parsing
- Error handling and recovery
Transcript Input → Context Building → AI Processing → Response Parsing → Structured Minutes
↓ ↓ ↓ ↓ ↓
[User Input] [Dictionary, Instructions, [API Call] [JSON Parsing] [UI Display]
Samples, Meeting Context]
- AIService: Main service class for AI interactions
- ApiManager: UI for provider and model configuration
- WorkspaceLayout: Three-panel editing interface
- SetupWizard: Guided onboarding flow
- Product Requirements Document - Complete feature specifications
- AI Pipeline Documentation - Technical implementation details
- CSS Design Reference - Design system guidelines
- Local Storage: All data stored locally in browser
- API Key Security: Keys encrypted and never sent to third parties
- Data Privacy: No data sent to external servers except chosen AI provider
- HTTPS Required: Secure connections enforced
- Start the application
- Click "Setup Wizard"
- Paste your meeting transcript
- Configure AI settings if needed
- Generate professional minutes
- Upload sample minutes to match your organization's style
- Create custom dictionaries for technical terms
- Set specific instructions for formatting preferences
- Export minutes in Markdown format
"AI service not configured"
- Ensure API key is entered in Settings
- Test connection to verify key validity
"CORS error"
- Use HTTPS or development server
- Check provider's CORS policies
"Rate limit exceeded"
- Wait before retrying
- Check provider's rate limits
See AI Pipeline Documentation for detailed troubleshooting.
This project follows standard development practices:
- TypeScript for type safety
- React with modern hooks
- Tailwind CSS for styling
- Comprehensive error handling
See LICENSE file for details.
Built with modern web technologies for professional meeting documentation.