Modern healthcare and service industries face significant challenges in managing appointments efficiently. Traditional booking systems are often manual, time-consuming, and prone to errors or double-bookings. Patients and clients expect seamless, real-time interactions—preferably through natural language and voice—while providers need intelligent scheduling that adapts to their availability and priorities.
Medora AI addresses these needs by providing:
- Frictionless, voice-first appointment booking for patients and clients.
- Intelligent, multi-agent orchestration that optimizes scheduling, resolves conflicts, and adapts to urgency.
- Automated notifications and reminders to reduce no-shows and improve communication.
- A scalable, extensible platform ready for future AI advancements.
- Conversational Booking: Users can book, cancel, or check appointments using natural language—via chat or voice.
- Voice Interaction: Speak to the system and hear responses back, making the experience accessible and hands-free. (Now powered by
streamlit-audiorecfor reliable browser-based recording.) - Multi-Agent Intelligence: Three specialized AI agents (User Bot, Doctor Bot, Scheduler Bot) collaborate to understand user intent, recommend doctors, and manage scheduling.
- Hybrid LLM/Intent Architecture: Routine actions (like booking/canceling) are handled by the orchestrator for reliability; all other queries are answered directly by a powerful LLM (default:
llama3-8b-8192on Groq). - Smart Scheduling: The system checks doctor/provider availability, validates time slots, and prevents double-booking.
- Conflict Resolution: If a requested slot is unavailable, the system suggests alternatives in real time.
- Priority Handling: Urgent and emergency requests are prioritized using a Model Context Protocol (MCP).
- Email Notifications: Users receive booking confirmations, cancellations, and reminders automatically.
- Dashboard & Manual Controls: Staff can view, add, or cancel appointments directly from the dashboard.
- Doctor Directory: Users can browse available doctors, their specialties, and schedules.
- Python 3.8+
- Streamlit: For the interactive web UI.
- LangChain & LangGraph: For multi-agent orchestration and conversational AI.
- Groq LLM (
llama3-8b-8192): For advanced language understanding and generation. - gTTS, pygame, SpeechRecognition: For text-to-speech and speech-to-text capabilities.
- streamlit-audiorec: For browser-based audio recording.
- smtplib, email: For sending email notifications and reminders.
- dotenv, yaml: For configuration management.
- Logging: For robust monitoring and debugging.
- pytz, numpy, sounddevice, vosk: For time zone handling and audio processing.
- Manual Scheduling Hassles: Eliminates the need for phone calls or manual entry by enabling conversational, automated booking.
- Double-Booking & Conflicts: Prevents scheduling errors by checking provider availability and resolving conflicts dynamically.
- Missed Appointments: Reduces no-shows with automated reminders and confirmations.
- Accessibility: Makes appointment management easier for users with disabilities or those who prefer voice interaction.
- Urgency Management: Ensures that urgent and emergency cases are prioritized appropriately.
- Scalability: Supports multiple providers and can be extended to other industries (consultants, salons, etc.).
- Staff Efficiency: Frees up administrative staff from repetitive scheduling tasks.
While Medora AI already leverages a robust multi-agent system, integrating an Agent Development Kit (ADK) in the future will unlock even greater capabilities:
- Continuous Learning: Agents can adapt to user preferences, learn from historical booking patterns, and personalize recommendations.
- Custom Agent Creation: Easily add new agents for specialized tasks (e.g., insurance verification, follow-up scheduling).
- Plug-and-Play Upgrades: Swap or upgrade agent logic without rewriting the core system.
- Advanced Analytics: Use ADK to analyze appointment trends, provider utilization, and user satisfaction.
- Cross-Domain Expansion: Extend the system to other domains (legal, education, wellness) by developing new agent modules.
- Python 3.8 or higher
- Streamlit
- Required Python packages (see
requirements.txt)
git clone https://github.com/extremecoder-rgb/medoraAI.git
cd medora-ai
pip install -r requirements.txt
streamlit run app.py- Book an Appointment:
Start a conversation with "I would like to book an appointment" or use the voice input (mic button next to chat input). - Check Availability:
Ask "Show me available doctors" or "What are the next available appointments?" - Manage Appointments:
View, cancel, or add appointments via the dashboard or chat. - Receive Notifications:
Get email confirmations and reminders automatically.
- Make sure your browser has microphone access enabled.
- If you see "Voice processing failed", try speaking clearly and for at least 2 seconds.
- Ensure you have a stable internet connection (Google Speech Recognition API is used).
- If you see a detailed error, check the stack trace for audio format or permission issues.
- If you see errors like
model_not_foundor 404, update your.envorsettings.yamlto use a supported model (e.g.,llama3-8b-8192). - Make sure your
GROQ_API_KEYis correct and you have access to the selected model. - Restart the app after changing model settings.
- Fork the repository
- Create your feature branch (
git checkout -b feature/AmazingFeature) - Commit your changes (
git commit -m 'Add some AmazingFeature') - Push to the branch (
git push origin feature/AmazingFeature) - Open a Pull Request
This project is licensed under the Apache License 2.0 - see the LICENSE file for details.
- Built with Streamlit
- Powered by Groq LLM (
llama3-8b-8192) - Voice processing by PyAudio, gTTS, SpeechRecognition, and streamlit-audiorec
- Multi-agent architecture inspired by LangChain
For support, email hsuswiowkskow@gmail.com or open an issue in the repository.
Made with ❤️ by Subhranil Mondal