AIdeas: Aether Health
Discover how our AI-powered platform reimagines the last mile of healthcare through the five elements of nature. By transforming complex clinical data into a calm, inclusive experience, we provide verified tools to combat health anxiety, decipher prescriptions, and detect counterfeit drugs.
App category
Social Impact
My vision
The last mile of healthcare is often broken, leaving patients feeling anxious, overwhelmed, and vulnerable to misinformation. Our team, Secret Society, wanted to reimagine this experience. We built Aether Health, an AI-powered healthcare platform that transforms complex clinical data into a calm, inclusive, and holistic experience inspired by the five elements of nature.
Unlike clinical apps that can induce panic, Aether Health is grounded in verified data and designed to solve critical everyday issues through its core elemental features:
- 🔥 VitalScan (Fire Element): An empathetic symptom checker using sentiment analysis and RAG architecture for anxiety-aware, verified assessments.
- 💧 HydroGuard (Water Element): An interactive map for water safety allowing users to report contamination via GPS-tagged photos.
- 💨 ClearScript (Air Element): An OCR tool that deciphers illegible doctor prescriptions, corrects drug names, and sets dosage reminders.
- 🌍 TrueMeds (Earth Element): A visual scanner that detects counterfeit drugs by analyzing packaging and pill attributes like color and shape.
- 🌌 LifeLoop (Space Element): The holistic layer connecting all elements with real-time translation and lifestyle advice.
Why this matters
Aether Health was built to protect and empower those who need it most. It bridges the gap between complex medical information and everyday human needs.
Beneficiaries:
- Anxious Patients: Receive calm assessments, avoiding panic-inducing internet search results.
- Vulnerable & Elderly Groups: Rural populations, the elderly, and non-English speakers gain accessible, life-saving tools. Patients who struggle to read doctor handwriting or complex medical labels can easily manage their care.
- Governments: Gain access to real-time citizen reports on local water contamination.
Problems Solved & Opportunities Created: Misinformation is a massive threat; Aether Health grounds its AI in WHO guidelines to prevent false self-diagnoses. It deciphers illegible prescriptions to prevent dangerous medication errors, and tackles the crisis of fake drugs (which affects 20% of developing markets) by detecting counterfeits. Ultimately, it empowers locals through citizen science (e.g., reporting sewage leaks), merges modern medicine with traditional wellness, and demonstrates that AI can be safe, empathetic, and transparent in critical care.
How I built this
To make this application scalable and accessible, we built Aether Health by maximizing Kiro and the AWS Free Tier.
Phase 1: Setup & Foundation We initialized the project using AWS Amplify to handle frontend hosting, CI/CD pipelines, and secure user authentication via Amazon Cognito. For UI implementation, user data, and storage, we utilized Amazon DynamoDB and Amazon S3.
Phase 2: Core Intelligence (Fire and Air)
- VitalScan: We implemented the symptom checker using Amazon Bedrock to access Nova Lite. We connected it to a Knowledge Base RAG containing WHO clinical guidelines for verified, anxiety-aware answers.
- ClearScript: Leveraging the superior multimodal capabilities of Nova via Bedrock, we built a tool to directly analyze images of handwritten prescriptions, extract text, and correct drug names in a single pass.
Phase 3: Visual and Map Features (Earth and Water)
- TrueMeds: We used SageMaker Studio Lab to train a lightweight computer vision model on a dataset of real vs. fake pill images. For our demo, we ran inference inside the notebook to keep it cost-effective.
- HydroGuard: We used Strands for geospatial intelligence. To maximize the free tier, we bypassed live government APIs and mocked the data by uploading a CSV of Safe/Unsafe Zones to S3, having the Strands agent query this static data to display pins on the interactive map.
Phase 4: Orchestration (Space)
- LifeLoop: We used Kiro to configure Bedrock AgentCore as the Supervisor Agent. Kiro scripted the logic to seamlessly route user intent—directing health questions to Fire and location queries to Water—ensuring a unified, holistic experience.
Demo
Watch Aether Health in action:
Here is a look at ClearScript analyzing a handwritten prescription
Here is a look of True Meds Feature:--
What I learned
Building Aether Health was a profound technical and emotional journey. Here are our key insights:
- Empathy in AI is Possible: By incorporating sentiment analysis and strict WHO guardrails in our RAG architecture, we learned that AI doesn't have to be cold and clinical. It can assess symptoms while actively de-escalating user anxiety.
- Orchestration is Key: Managing five distinct features could have resulted in a fragmented app. Using Kiro to configure Bedrock AgentCore as a Supervisor Agent taught us how crucial intelligent routing is for a seamless user experience.
- Doing More with Less: We learned how to stretch the AWS Free Tier to its absolute limits. By getting creative—like mocking geospatial data for Strands via S3 and running lightweight CV models in SageMaker Studio Lab—we proved you can build a highly sophisticated, multi-modal application without breaking the bank.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article