GRADGRABS

GradGrabs helps new grads claim free gifts from major brands like Chick-fil-A, LEGO & Rare Beauty. AI writes the perfect letter for each one in seconds ready to email or print and mail.

Comment

Inspiration

It started with a TikTok rabbit hole. Thousands of graduates were mailing their announcement cards to brand headquarters and receiving boxes of free merch back — Chick-fil-A sending $150 gift boxes, LEGO mailing personalized sets, Rare Beauty dropping full product kits. Response rates were surprisingly high, sometimes over 80%.

But the process was completely manual. You had to find the right mailing address yourself, write a letter from scratch, and repeat it one brand at a time. Most graduates either never heard about it or gave up after the first attempt. That gap felt like a product.

What it does

GradGrabs turns your graduation announcement into a personalized outreach campaign across 28 verified brands in under two minutes.

You enter your name, degree, and graduation year. The app shows every brand worth contacting, sorted by response rate and confidence level. Pick your targets, hit generate, and AI writes a tailored letter for each one , referencing the specific brand, your degree, and your personal story.

Each letter works two ways: paste it into the brand's contact form online, or print it and mail it with your physical announcement card. A built-in tracker lets you log what you've sent, what's pending, and what arrived.

How I built it

  • Frontend: React + Vite, CSS Modules, dark minimal UI
  • AI layer: Gemini API for personalized letter generation via a Vercel serverless proxy
  • Data: 28 manually verified companies with real contact form URLs and physical mailing addresses, sourced from 2025–2026 graduation season reports
  • Logos: Hunter Logo API with emoji fallbacks for smaller brands
  • Storage: Fully client-side — localStorage for profile, selections, generated letters, and tracker state. No backend, no database, no accounts required

Challenges I ran into

Making letters feel human. Early generations were generic and templated. The fix was tightening the prompt — constraining length, banning specific clichés like "I hope this finds you well", and injecting the student's personal story as context. The quality jump was immediate.

Token bleed from Gemini. During early testing with Gemini 3.5 Flash, responses would cut off mid-sentence and append raw metadata strings like MAX_TOKENSTEXTstandard. Fixed by bumping maxOutputTokens, stripping garbage characters from responses. Still using the free tier because I'm broke though.

Data integrity. Most "graduation gift" resources online are outdated or fabricated. We cross-referenced every company against actual 2025–2026 TikTok and Reddit reports before including them. Several popular sources listed fake email addresses — i replaced all of them with verified contact form URLs.

Scope discipline. The temptation was to build a live agentic system that dynamically crawls the web for companies. I prototyped it but cut it for submission — the curated static list with high-quality AI letters was more reliable and more useful than a live system returning noisy results.

Accomplishments that I'm proud of

  • 28 verified brands with accurate addresses and real response rate data — no fabricated contacts
  • Letters that pass the "did a human write this" test, with each one genuinely referencing the specific brand
  • Zero backend required — the entire user journey including letter generation, tracking, and persistence works client-side
  • A print-ready letter flow that bridges digital and physical outreach in one click
  • Company logos via hunter with graceful emoji fallbacks, keeping the UI clean even for niche brands

What I learned

A well-scoped product with real data beats an over-engineered one with impressive architecture. The most valuable part of GradGrabs isn't the AI , it's the 28 verified brands with accurate addresses and realistic response rates. The AI makes it fast. The data makes it trustworthy. Also worked with react after a really long time found out i'm very rusty.

What's next for GRADGRABS

  • Live company discovery using web search APIs to surface new brands each graduation season automatically
  • Response tracking with photos — upload a picture of what arrived and build a public wall of wins
  • Graduation season calendar — brands respond faster in May–June; smart send timing recommendations
  • Email integration — one-click send directly from the app via Gmail OAuth, no copy-paste needed
  • Expand beyond the US — source equivalent brands and addresses for UK, Canada, India, and Australia graduation markets

Built With

  • gemini-api
  • hunter.io-api
  • react+vite
Share this project:

Updates

Submission history