Inspiration

Most people are not prepared for critical situation that take place all of a sudden, and in the moment, even simple decisions feel hard. We noticed the same pattern: people panic, freeze, or start searching random things over internet while precious minutes pass. We built First30.ai to be the calm guide in that gap, the first 30 minutes before professional help takes over.

The idea is not to replace doctors or paramedics. It is to help a normal person do the next right thing without feeling lost.

What it does

First30.ai is a voice guided emergency helper for everyday people.

You choose what situation is happening, and it walks you through clear steps one at a time
It speaks the steps out loud, so you can keep your hands and eyes on the situation
It can show nearby hospitals, pharmacies, and emergency services based on your location
At the end, it generates a short summary you can share with responders, like what you observed and what you already did

Target Group

Primary: General public with no medical training (bystanders, family members)
Secondary: First responders, emergency dispatchers, community health workers
Tertiary: Healthcare organizations, insurance companies, public health agencies

Features & Functions

15+ emergency category flows with step-by-step guidance
ElevenLabs-powered voice narration in 5 languages
Location-based nearby hospital/pharmacy/emergency services finder
Emergency summary generator for handoff to responders
Dispatcher dashboard prototype for structured triage
Accessibility mode with large text and auto-voice

Value Proposition & USP

Traditional First Aid	First30.ai
Requires training or prior knowledge	Built for zero training
Panic often leads to freezing	Calm, one step at a time guidance
Language barriers limit usability	Voice guidance in 5 languages
Users waste time searching what to do	Nearby help is located automatically
No clear handoff to responders	Generates a quick summary for responders

User Feedback

"I showed this to my elderly mother. She said for the first time she feels like she could actually help in an emergency instead of just panicking." — Beta Tester, Munich
"The voice feature is a game-changer. I don't have to look at my phone while doing CPR." — Nurse, Hamburg
"Finally something that works in Hindi. My parents can't read English emergency guides." — Software Engineer, Delhi

Business Model

Revenue Stream	Description
B2B Licensing	White label for hospitals, insurance companies, and corporate wellness programs
Freemium	Core features free; premium voices, offline packs, and family accounts
Public Health Grants	Government and NGO funding for public deployment
API Access	Emergency guidance API for third party apps and platforms
Training Partnerships	Integration with first aid certification and training programs

Implementation & Feasibility

Current Status: Working prototype with all core features functional Next Steps:

Medical content review by certified professionals (2 weeks)
User testing with 50+ participants (1 month)
Offline mode implementation (2 months)
Pilot with 2–3 healthcare partners (3 months)
Public beta launch (4 months)

Data Requirements + GDPR/DSGVO Compliance

Data Handling

Data Type	Collection	Storage	Purpose
Location	On request	Not stored	Nearby services lookup
Language preference	Local setting	Browser only	UI personalization
Emergency session	Temporary	RAM only	Summary generation

GDPR Compliance

✅ No personal data collection
✅ No user accounts required
✅ No cookies beyond essential functionality
✅ Processing happens client side or in ephemeral edge functions
✅ Clear privacy policy and consent dialogs

How we built it

We built the app with React, and TypeScript so it is fast and easy to maintain.

The emergency guidance is data driven. Each emergency category is a structured flow made of steps and safety notes. The UI reads that data and renders the experience, which makes it easy to add new categories without rewriting the app.

For voice, we use ElevenLabs. To keep the API key secure, the app calls a Supabase Edge Function. The browser sends only the text, the function calls ElevenLabs using the secret key, and the audio is returned to the app.

For nearby help, we use location only when the user asks for it, then show nearby services and quick directions.

Challenges we ran into

Designing and choosing the features were the challenge that we faced. Too many options can make people confusing.
Location features needed to work even when permissions are denied or the network is unstable.
Writing step by step flows that are clear and safe across many emergency types was harder than expected.
Making everything accessible without making the UI feel heavy or cluttered.

Accomplishments that we're proud of

A working prototype with 15+ emergency flows that stays consistent and easy to follow
Voice guidance that users can follow without constantly looking at the screen
A clean emergency summary that supports better handoff to responders
A scalable design where adding a new emergency is mostly adding data, not building new screens
A privacy first approach with no accounts and minimal data handling

What we learned

In emergencies, clarity beats features. One good next step is better than a long guide.
Voice is not just a nice add on, it changes how usable the app is under stress.
Data driven design makes the project easier to grow and maintain.
Privacy and security decisions affect trust, so we treated them as core requirements.