Inspiration
Rural areas in Canada often have very limited access to doctors. That lack of care inspired me to create a browser-based AI assistant that offers basic health guidance.
What it does
HealthVoice lets users speak their symptoms in a browser. It uses the browser’s SpeechRecognition API to capture speech. That text is sent to Amazon Bedrock, which returns an AI-generated health advice. The advice appears as plain text in the UI.
How we built it
- Voice input: Used browser’s SpeechRecognition API to record spoken questions.
- AI processing: Transcribed text is sent to AWS Bedrock to generate a simple health response.
- Displaying result: Bedrock’s reply is returned and shown to the user as text.
Challenges we ran into
- AI safety: Ensuring responses are clear and effective.
Accomplishments that we're proud of
- Delivered a real helper for rural users without needing high‑bandwidth connections or video calls.
- Created a browser-based voice-first health assistant powered only by speech recognition and AWS Bedrock, minimising service cost of AWS services.
What we learned
- How to capture and process voice input in browser using the Web Speech API.
- Better understanding of AWS and how to use Bedrock Models.
What's next for Health Voice AI
- Feedback button: Let users indicate if the advice helped.
Built With
- amazon-bedrock
- amazon-web-services
- css
- express.js
- html
- javascript
- node.js
Log in or sign up for Devpost to join the conversation.