OfferShield is an AI-for-Good platform that helps students, early-career professionals, and job seekers instantly verify whether a job or internship offer letter is real or a scam. Users can upload a PDF/image or paste text; the system extracts content with OCR, analyzes it with multi-layer AI checks, flags suspicious phrases, evaluates salary realism, detects scam patterns, and provides a transparent explanation.
Scams targeting students are increasing globally — fake HR emails, deposit-based joining scams, data-entry fraud, identity theft. Many victims don’t know how to verify offer letters. OfferShield fills this trust gap by giving them a simple, fast, accessible online tool requiring no installation, no registration, and no expertise.
The system includes:
Client-side AI analysis (OpenAI mini + custom heuristics)
Cloudflare Worker backend for secure API key protection
Firebase real-time community reporting
OCR pipeline (PDF.js + Tesseract)
Red-flag highlighter
Salary realism scoring
Link reputation analysis
Safer corporate-style offer rewrite
1-click community scam report fingerprinting
The UI is built mobile-first and optimized for usability. OfferShield directly contributes to “AI for Good” by protecting vulnerable populations from fraud, financial loss, and identity theft.
This week’s new progress includes: full Cloudflare Worker deployment, scam-pattern matching engine, red-flag highlighting system, new UI, safe rewrite generator, and public demo link.
✅ PROBLEM STATEMENT (Pain Point)
Millions of students and job seekers receive fake job offer letters demanding deposits, training fees, or personal data. These scams cause:
Financial loss
Identity theft
Emotional stress
Broken trust in legitimate recruiters
There is no instant, accessible online tool that can reliably check an offer letter.
OfferShield solves this.
✅ WHY IT’S AWESOME (Judges love this section)
OfferShield stands out because it is:
⭐ Instant
Paste text or upload a file → get a complete scam-risk breakdown in seconds.
⭐ AI-Driven
Uses OpenAI mini + custom NLP heuristics to detect fraud patterns.
⭐ Privacy-Safe
Keys are protected using Cloudflare Workers + Secrets Store.
⭐ Community-Powered
Users anonymously report scam offers → fingerprinted → displayed to everyone.
⭐ Language-Agnostic
Works on English, Indian English, basic Chinese, and any language that OCR can handle.
⭐ Real-World Impact
Students frequently fall for fake internship/job offers — this tool prevents that.
⭐ Hackathon-Friendly
Runs fully in the browser; no heavy backend; deploys anywhere.
✅ NEW PROGRESS MADE THIS WEEK (required for old projects)
In the last 7 days, I added:
Cloudflare Worker backend + secret-protected API
New UI + animations + trust meter
Red-flag phrase highlighter
Scam similarity score engine
Salary realism & link reputation module
Community-report Firebase integration
“Generate Safe Version” AI rewrite
Improved OCR (PDF.js + Tesseract)
Performance upgrades and code restructure
Public demo deployment
README + documentation updates
This satisfies the hackathon rule: any new progress counts.
Log in or sign up for Devpost to join the conversation.