ToS Summarizer
A Chrome extension that summarizes Terms of Service and Privacy Policies with AI — and rates the risk before you accept.
It all started with a post by Pedro Fonseca: what if AI summarized Terms of Service? The idea stuck — why not bring it into the browser, one click away? There were two challenges. The user's: nobody reads ToS (long, dense, legalese) and we accept clauses about our own data blindly. And mine, personal: this was my first project built entirely with AI, and I wanted to use it to prove I could ship a complete end-to-end integration — payments, secure key management and a real production deploy.
The MVP came together fast, with Cursor as co-pilot. But I soon saw the real adoption blocker: forcing each person to paste their own API key was a barrier. The goal became twofold — accessible to anyone and, above all, secure.
The “experiment” became my biggest project to date. I built the infrastructure around it: a full Vercel deploy, credit-based payments with Stripe, a secure API-key management system (the key never leaves the server) and a management dashboard. Analysis is done by Google Gemini; history and accounts live in a database.
For any ToS or policy, it returns a structured summary (overview, key points, privacy alerts) and rates risk, complexity and best practices from 1 to 10, with traffic-light indicators. You choose the focus (privacy, terms or balanced), keep a filterable history and export to JSON/CSV/TXT. A Stripe credit system with secure key management lets you use the shared API — without each person configuring their own — or your own Gemini key. All topped by a management dashboard.
More than adoption, this project proved capability — but adoption showed up too. Published on the Chrome Web Store, the extension has around 356 installs with only 3 uninstalls: near-total retention. It was my first product built entirely with AI and the first time I wired up, end to end, payments (Stripe), secure API-key management and a production deploy on Vercel. The method that came out of it — validate the idea, build with AI as co-pilot and ship to production safely — became the foundation of everything I built afterwards.