AI Mental Health Companion Agent

• 

  Loving the daily affirmations and mindful breathing reminders from Mindful Companion.

Inspiration Mental health care is often inaccessible due to stigma, high costs, or lack of available professionals. With AI, we saw an opportunity to create a safe, empathetic, and always-available companion that supports emotional well-being while ensuring responsible escalation when users need real help.

What it does Acts as a 24/7 mental health companion through text/voice.

Provides daily mood check-ins and journaling.

Suggests guided meditation, breathing, and CBT-based exercises.

Uses sentiment analysis to detect distress or emotional shifts.

Escalates to mental health professionals or helplines when necessary.

How we built it Amazon Bedrock → for natural, multilingual conversations.

Amazon Comprehend → for emotion and sentiment analysis.

AWS Lambda + Step Functions → to orchestrate check-ins, analysis, and escalation workflows.

Amazon DynamoDB → to securely store anonymized journaling and user progress.

Amazon Polly → to deliver soothing, human-like voice interactions.

Amazon SNS → for sending alerts or connecting users to professionals.

Challenges we ran into Designing AI conversations that feel empathetic, not robotic.

Maintaining privacy and security while storing user data.

Handling multilingual and culturally sensitive interactions.

Striking a balance between automation and human escalation.

Accomplishments that we're proud of Built a working prototype that combines empathy, analysis, and escalation.

Created a multi-agent system (Companion Agent, Analysis Agent, Escalation Agent).

Designed a solution with the potential for global social impact.

What we learned How to build human-centered AI agents that prioritize trust and empathy.

Best practices for combining AWS AI services into real-time applications.

The importance of responsible AI design in sensitive areas like mental health.

What's next for AI Mental Health Companion Agent

Collaborate with mental health professionals to improve accuracy and trustworthiness.

Add wearable integration (heart rate, sleep, activity) for deeper insights.

Develop offline-first capability for areas with poor connectivity.

Expand into a mobile app with broader accessibility and personalization.

Built With

• animations
• class-variance-authority
• color
• css
• custom
• hsl
• lucide
• react
• responsive
• router
• shadcn/ui
• tailwind
• tanstack
• typescript
• vite