Suraksha Sathi

An ASI-1 powered scam and phishing analyzer that helps Indian users detect fraudulent messages, links, and OTP traps, then get clear next-step guidance in seconds

Comment

Inspiration

Online scams in India have become extremely common, especially fake bank alerts, KYC expiry messages, UPI fraud, OTP theft, parcel scams, and impersonation attacks. Many people can recognize obvious spam, but highly urgent and convincing scam messages still create panic and lead users to share sensitive information.

We built Suraksha Sathi to make scam detection fast, practical, and easy to understand. Instead of giving users a generic chatbot response, we wanted to build a tool that could inspect suspicious content, explain exactly why it looks dangerous, and provide immediate next steps in plain language.

What it does

Suraksha Sathi is an ASI-1 powered scam and phishing analyzer built for Indian users.

A user can paste a suspicious SMS, email, offer, payment request, or phishing link into the app. The system then uses ASI-1 in two stages:

  1. It performs a structured threat analysis using schema-constrained output.
  2. It uses that structured result in a second ASI-1 pass to generate clear next-step guidance.

The result is a full risk report with:

  • verdict
  • risk score
  • attack type
  • confidence
  • evidence and red flags
  • immediate actions
  • possible data exposure
  • plain-language safety guidance

This helps users quickly understand whether a message is likely fraudulent and what they should do next.

How we built it

We built the project as a lightweight web app with:

  • a Node.js backend
  • a vanilla HTML, CSS, and JavaScript frontend
  • ASI-1 Chat Completions API for analysis and guidance
  • Vercel for deployment
  • GitHub for source control and submission

The backend accepts suspicious text through /api/analyze, sends it to ASI-1 with a strict JSON schema, and receives a machine-readable scam report. That report is then passed into a second ASI-1 prompt that converts the analysis into practical and shareable user guidance.

We also added normalization logic so the app can handle imperfect model outputs and still present sensible risk scores, verdicts, and confidence values.

Challenges we ran into

One of the biggest challenges was making the app feel like a real cybersecurity product rather than just a prompt wrapper.

We wanted the ASI-1 integration to be meaningful, so we designed a two-step reasoning flow:

  • first for structured extraction
  • then for actionable remediation guidance

We also had to solve practical implementation issues:

  • making the frontend render model output cleanly
  • handling inconsistent score/value outputs safely
  • making the project work both locally and on Vercel
  • restructuring the backend into Vercel-compatible API routes for deployment

Another challenge was keeping the project simple enough to finish quickly while still making it credible for judges.

What we learned

We learned a lot about building reliable AI-powered applications beyond basic text generation.

Key learnings included:

  • how to use ASI-1 for structured, schema-based outputs
  • how multi-step prompting can improve clarity and usefulness
  • how to design safer AI UX for cybersecurity scenarios
  • how to normalize model outputs to improve product reliability
  • how to adapt a local Node app into a deployable serverless architecture

Accomplishments that we're proud of

We are proud that Suraksha Sathi:

  • uses ASI-1 as a genuine reasoning engine, not just a chatbot
  • addresses a real and urgent problem affecting Indian users
  • produces both technical analysis and plain-language guidance
  • is fully deployed and submission-ready
  • has a simple interface that judges can understand quickly

What's next for Suraksha Sathi

We would like to extend the project with:

  • Hindi and Hinglish support
  • screenshot and OCR-based scam detection
  • URL reputation checks
  • scam category analytics
  • quick reporting options for banks and cybercrime portals
  • user education flows for vulnerable first-time users

Suraksha Sathi is designed as a practical starting point for AI-assisted scam defense, and we believe it can grow into a useful real-world safety tool.

Built With

Share this project:

Updates