AI Clinic Assistant → M Clinic
Introduction
The AI Clinic Assistant (now M Clinic) is a fully operational, AI-powered telephone receptionist designed to automate appointment handling in healthcare clinics. It conducts real-time, natural voice conversations with patients using OpenAI's GPT-4o Realtime API — handling bookings, rescheduling, and cancellations automatically.
Originally prototyped using landline and Android-based approaches (V1–V9), the system has since evolved into a cloud-based VoIP platform with dedicated UK phone numbers, real-time Google Calendar sync, SMS confirmations, and a customer management panel.
The project was first piloted in a private clinic in London, where it received positive feedback from reception staff and clinicians — validating its real-world applicability and practical value.
Problem It Solves
Many small clinics and solo practitioners lose time, patients, and revenue by handling appointment-related calls manually. Staff are overwhelmed with repetitive tasks during busy hours, and missed calls mean missed patients — especially for doctors who work alone without a receptionist.
M Clinic fills this gap by offering an easy-to-deploy, affordable, and reliable AI receptionist that works 24/7.
Key Features
Natural voice interaction: Understands and responds to patients like a real receptionist
Dedicated UK phone number: Each clinic gets its own number — no hardware required
Real-time Google Calendar sync: Appointments appear instantly in the doctor's calendar
SMS confirmations: Patients receive automatic text messages for bookings, changes, and cancellations
Customer panel: Clinic owners manage doctors, schedules, and appointments through a web dashboard
Zero technical setup: Clinics can start using it without IT teams or programming knowledge
Scalable design: Adaptable for GP clinics, dental practices, beauty clinics, therapy centres, etc.
24/7 coverage: Handles calls outside working hours, during lunch breaks, and on weekends
Project Status
Successfully tested in a real clinical setting in London
Refined through feedback from clinic managers, receptionists, and doctors
9 versions developed — from basic voice prototype to full commercial product
Now commercially available for private clinics at https://mclinic.barisy.org
Screenshots & Demo Video
▶️ Watch demo videos at https://mclinic.barisy.org
GitHub Repositories (Early Open-Source Versions)
Note: These repositories contain the early open-source versions of the project, shared to document the development journey and contribute to the community. The current production system (M Clinic) has evolved significantly beyond these versions and is no longer open-source. These early versions remain publicly available as a reference for anyone interested in the technical foundations.
clinic-assistant-v1: (March 2025) Basic logic for voice-to-text transcription and Google Calendar syncing.
clinic-assistant-v2: (April 2025) Enhanced error handling, improved name/date/time extraction, and more stable appointment creation.
clinic-assistant-v7: (July 2025) Final polished version used in the public demo — supports full voice interaction with GPT, natural flow tracking, and appointment confirmations.
clinic-assistant-v8: (August 2025) Plug-and-play USB modem integration for real-time landline call detection and automatic voice response — ready for deployment in live clinic environments.
clinic-assistant-v9: (October 2025) Mobile-first Android release with real-time auto-answering, natural GPT-powered conversations, and direct Google Calendar integration — removes PC and hardware dependencies.
What Was Planned vs What Was Built
Many of the features originally listed as "planned" have now been implemented in the commercial version of M Clinic:
✅ Interactive Web Dashboard → Customer Panel (live)
✅ Customizable Voice & Greeting → Configurable per clinic via panel
✅ Google Calendar Sync → Real-time two-way sync (live)
✅ Secure API Key Handling → Server-side, not exposed to users
✅ User Account Management → Google sign-in with clinic-specific settings
✅ Usage-Based Billing → Usage tracking with minutes and SMS counts
✅ Cloud-Native Architecture → Firebase + Hetzner VPS deployment
Timeline
This page presents the development timeline of the AI Clinic Assistant and related innovation milestones.
From the initial idea sparked by real-world problems in a clinic, to working prototypes, public showcases, and a commercial launch — each phase reflects a hands-on, purpose-driven approach to building scalable, human-friendly AI systems.