SafeReach AI

About the Project

SafeReach AI was inspired by a simple but serious problem: during emergencies, vulnerable people often face two dangers at the same time — the actual crisis and scams that exploit fear, urgency, and confusion.

A person may receive a message about flood relief, bank verification, hospital payment, government aid, emergency donation, or a family member in danger. In stressful moments, it can be difficult to know what is real, what is risky, and what action to take next.

SafeReach AI was built to help users answer one urgent question:

Is this message risky, and what should I do safely right now?

What I Built

SafeReach AI is a global no-login safety assistant for seniors, families, caregivers, migrants, travelers, and vulnerable users.

The app includes:

• Live Safety Mode for real user-entered messages and crisis guidance
• Demo Mode for sample scenarios and hackathon walkthroughs
• Scam-risk message analyzer
• Guided safety chat assistant
• Crisis navigator with structured safety steps
• Senior-friendly mode with large buttons
• Emergency wallet card generator
• Family alert message generator
• Self-help safety center
• Country safety packs
• Privacy-safe local analytics

The scam analyzer uses rule-based risk scoring. It looks for risk indicators such as OTP requests, bank/KYC language, suspicious links, urgent payment pressure, fake relief claims, fake charity requests, family emergency pressure, and instructions not to call anyone.

Instead of saying a message is definitely a scam, SafeReach AI uses safer wording:

“This message shows scam-risk indicators.”

This keeps the app responsible while still giving useful guidance.

Inspiration

The inspiration came from real-world situations where people, especially seniors and vulnerable users, are targeted during moments of panic.

Scammers often use messages like:

• “Your bank KYC is expired. Enter OTP now.”
• “Flood relief approved. Pay a processing fee.”
• “Your son is in hospital. Send money immediately.”
• “Government aid available. Upload your ID and bank details.”
• “Donate now to an official rescue team.”

These messages combine urgency, emotion, money, and confusion. SafeReach AI was designed to slow the user down, highlight red flags, and guide them toward safer action.

How I Built It

I built SafeReach AI using MeDo through prompt-driven development and iterative refinement.

MeDo helped turn a complex product idea into a working web app by allowing me to describe the product structure, user flows, safety rules, UI behavior, and demo scenarios step by step.

The build process included:

• Designing the Live Safety Mode and Demo Mode structure
• Creating a global country and language selector
• Building the suspicious message analyzer
• Adding rule-based scam-risk scoring
• Creating crisis guidance flows
• Designing a senior-friendly interface
• Adding a guided safety chat assistant
• Creating emergency wallet card and family alert tools
• Adding privacy-safe analytics
• Testing the app for mobile usability, disclaimers, and safe behavior

MeDo was useful because I could rapidly refine the app after testing. For example, I improved the mode structure so the app would not feel demo-only. Live Safety Mode became the default, while Demo Mode remained available for judges and walkthroughs.

What I Learned

I learned that safety apps must be designed differently from normal productivity apps. The interface must be simple, calm, and direct because users may be stressed, confused, elderly, or in danger.

I also learned the importance of clear boundaries. SafeReach AI does not contact emergency services, does not diagnose medical problems, does not give legal advice, and does not guarantee scam detection. Instead, it provides structured safety guidance and encourages users to verify through official sources or trusted people.

I also learned how powerful prompt-driven development can be when building a multi-feature product. MeDo helped me move quickly from idea to working interface while continuing to improve the product through testing.

Challenges I Faced

The biggest challenge was balancing usefulness with safety.

A scam and crisis app must be helpful, but it must not overclaim. I had to carefully design the language so the app says “risk indicators” instead of “confirmed scam.”

Another challenge was avoiding a demo-only feeling. The app originally focused heavily on demo scenarios, but I refined it so Live Safety Mode became the default and Demo Mode became optional.

A third challenge was making the app usable for seniors. That required large buttons, simple text, clear warnings, and direct actions like “I Am Safe,” “Do Not Share OTP/PIN,” and “I Received Money Request.”

Built With

• MeDo
• Prompt-driven app development
• Web app interface
• Rule-based risk scoring
• Responsive UI design
• Local app state
• Privacy-first no-login structure
• Guided safety chat logic
• Global country safety packs
• Demo and live mode workflows

What’s Next

Future improvements could include:

• Offline crisis safety packs
• More countries and languages
• Voice/audio mode for low-literacy users
• Verified local safety partners
• NGO/admin safety content dashboard
• Official emergency alert integrations
• Encrypted emergency card storage
• Caregiver sharing mode
• More advanced scam-pattern detection
• Accessibility improvements for seniors and disabled users

Final Summary

SafeReach AI is a global safety assistant that helps users detect scam-risk indicators, follow crisis guidance, use guided safety chat, generate family alerts, and create emergency wallet cards.

Using MeDo, I was able to transform a complex safety concept into a working no-login web app with Live Safety Mode, optional Demo Mode, senior-friendly design, and privacy-safe local analytics.

Built With

• medo