I’m incredibly excited to finally share what I’ve been building over the last few weeks:  Vera. 🚀 We've all been handed a 20-page contract or terms of service agreement and just blindly signed it because who has the time (or the legal degree) to read through the jargon? I built Vera to fix that. Vera is an AI-powered risk engine that instantly scans your contracts, highlights high-risk clauses, identifies financial exposure, and translates legal jargon into plain English. As a solo developer, building a product that handles sensitive user data required an absolute obsession with architecture, security, and user experience. Here’s a look under the hood at how I built it: Tech Stack: Next.js (App Router), TypeScript, Tailwind CSS AI Integration: Rather than just slapping a generic prompt into an LLM, I engineered a highly constrained, multi-stage reasoning pipeline. Vera utilizes DeepSeek/OpenAI strict structured outputs (JSON schemas) and specialized prompt architecture to deterministically extract legal red flags, financial exposure, and clause conflicts with zero hallucination. Security First: Implemented strict Content-Security-Policies (CSP), server-side Firebase session revocation, and robust API rate-limiting. Contracts are processed 100% in-memory, files are never stored on a hard drive or cloud bucket. Flexible Payments: Integrated LemonSqueezy to handle both $10/mo recurring subscriptions and flexible $5 one-time top-ups, synced seamlessly with a Firebase backend. Polished UX: Built custom React drag-and-drop zones, buttery smooth GPU-accelerated page transitions, and intelligent client-side routing guards to completely eliminate UI flash. Building a secure, production-ready SaaS involves so much more than just hooking up an API wrapper. Vera is the most robust web app I’ve built to date. I'd love for you to check it out! Let me know what you think of the UI and the analysis accuracy. 👇 https://verahq.xyz #BuildInPublic #Nextjs #SaaS #WebDevelopment #AI #CyberSecurity #TypeScript