Alira: Autism Care Companion

App Main Features
Dashboard + Encouragement Alert for undiagnosed child to schedule a visit to a clinic.
Modules (Age Filtered)
Modules Activity
Nearby Clinics
Progress Report
AI Land Page
AI Accuracy to answer important concerns and refuse irrelevant questions
Landing Page
Settings + Intergration of Bilingual System
AI-Powered ChatBot
Statistics
Sign-Up Interface
AI-Powered ChatBot

Inspiration Stigma survives in the dark. In the Philippines, the heaviest barrier to autism care isn't just the lack of specialized clinics—it's the silence. When a child develops differently, society often makes it feel unacceptable. Out of sheer love, parents build walls of denial to protect their families from a label. But those walls don't keep the judgment out; they only trap the child in the shadows.

We realized that the most powerful force in a child's life isn't the specialist they see for one hour a week—it is the parent they live with for the other 167 hours. We built ALIRA because the only cure for stigma is awareness. We wanted to give Filipino families the light to stop hiding, the tools to understand their child, and the courage to just be parents again.

What it does ALIRA (Autism Learning and Intervention Resource App) is a bilingual (English/Filipino) Autism Care Companion designed to hold a parent's hand in the dark. It is an awareness and support system featuring:

Learning Modules: Evidence-based coaching activities (grounded in DIR/Floortime, TEACCH, and NDBI) designed for home use to turn hours of helplessness into hours of progress.

Progress Tracking: Visualizes a child's developmental journey and activity scores over time using interactive Recharts.

ALI Chat: A conversational AI companion powered by the Manus LLM API, prompted to be culturally sensitive to Filipino families and to provide evidence-informed, compassionate guidance.

Nearby Clinics & QR Demographics: Uses Google Maps to locate specialized clinics. It features a unique QR Demographics Card that pre-loads the child's details, streamlining stressful clinic registration desks when families finally take that brave step to seek professional care.

How we built it We prioritized speed, reliability, and developer experience to build a highly stable MVP during the hackathon.

We built the Frontend using React 19, Tailwind CSS 4, and shadcn/ui for a clean, accessible, and fast interface.

Our Backend runs on Node.js and Express 4, utilizing tRPC 11.

For the Database, we used MySQL/TiDB managed through Drizzle ORM. This combination of tRPC and Drizzle provided strict end-to-end type safety.

AI & Storage were handled via the Manus Built-in Forge API (OpenAI-compatible) and AWS S3.

We implemented a custom i18n React Context that maps a typed dictionary across the UI, allowing instant, app-wide language switching between English and Tagalog without page reloads.

Challenges we ran into Our biggest challenge was balancing empathy with clinical safety. Because we are dealing with a sensitive medical condition, we had to strictly boundary our AI (ALI) to act purely as a coach. It is hard-coded to never diagnose and to firmly redirect clinical questions to a developmental pediatrician.

Technically, managing the Google Maps integration (JavaScript API, Places, Geocoding) without exposing our API keys on the frontend required us to successfully implement a proxy system. Additionally, designing a relational database schema capable of handling multiple child profiles, module scoring, and persistent bilingual preferences under strict time constraints heavily tested our architectural planning.