DEV Community

Ishant Singh
Ishant Singh

Posted on

SmartCarePlus: Voice Medicine Reminders Powered by Google Gemini

#devchallenge #googleaichallenge #ai #gemini

This is a submission for the Google AI Studio Multimodal Challenge

What I Built

I created a WhatsApp voice-call medicine reminder system.

Patients can receive automated WhatsApp calls that play a personalized voice message reminding them to take their medication. The audio is generated with Google Gemini 2.5 Flash Preview TTS and delivered through the WhatsApp Cloud API.

For now, it’s user-initiated so anyone can test it directly:

📞 Just call +91 94572 57934 on WhatsApp — you’ll hear the Gemini-generated reminder spoken back to you, and then the call will terminate automatically.

This demo shows how healthcare providers could make medication reminders more engaging, reliable, and accessible compared to simple text notifications.

Demo

  • 📹 Click to view demo
  • 🖼️ Screenshots of the WhatsApp call in progress
  • 🖼️ Screenshots of my FastAPI/Cloud Run service handling the webhook

How I Used Google AI Studio

  • I used Gemini 2.5 Flash Preview TTS from AI Studio to synthesize natural-sounding voice reminders.
  • My Cloud Run backend:
    • Receives WhatsApp call events (via webhook).
    • Calls Gemini TTS API with the reminder message.
    • Streams the generated WAV audio into the WhatsApp Cloud API call.
    • Terminates the call after playback.

For verification, I included:

  • Cloud Logging exports showing live requests to Gemini with timestamps and model name.
  • A sanitized Gemini response JSON showing the model field and request ID.
  • A SHA-256 hash of the demo audio file (which matches what plays during the call).

Multimodal Features

  • Text → Speech (Gemini TTS): Turns reminders into natural spoken audio.
  • Audio Playback via WhatsApp Calls: Patients hear the voice instead of only reading a message.
  • Real-world Integration: Combines AI voice with messaging infrastructure, reaching patients where they already are.

This multimodal experience enhances accessibility and increases the chance patients take their medication on time.

Future Use Cases

The current demo is a simplified proof-of-concept. In a real deployment:

  • The reminder would include the actual medicine name, dosage, and instructions.
  • The system could listen for the patient’s reply (e.g., confirm they took the medicine, request a reschedule, or ask a question).
  • It would have secure access to the patient’s health data stored in SmartCarePlus (my healthcare platform) to personalize reminders and update adherence records.
  • Doctors and caregivers would see adherence status in real time, improving transparency and outcomes.

Proof of Gemini Usage

  • 🖼️ Cloud Logging screenshot: Vertex AI request log with model gemini-2.5-flash-preview-tts.
  • 🖼️ AI Studio usage dashboard screenshot (yesterday’s usage entry).

Top comments (0)