The latest articles on DEV Community by xQrewo (@xqrewo). https://dev.to/xqrewo https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3176658%2Ffc442fd0-0e47-4e55-8f23-0fc09f1bbf17.jpeg https://dev.to/xqrewo en xQrewo Thu, 29 May 2025 06:22:46 +0000 https://dev.to/xqrewo/building-docuscopeai-my-first-real-full-stack-ai-project-part-1-3k7p https://dev.to/xqrewo/building-docuscopeai-my-first-real-full-stack-ai-project-part-1-3k7p <p><strong>Hey everyone!</strong><br> I'm Max, a high school student from Poland. I’ve been coding for a while now, but this is the first time I’m diving deep into a full-stack project that mixes frontend, backend, and AI magic — and I'm hyped to share my journey with y’all. This post is the first in a series documenting how I’m building DocuScopeAI from scratch.</p> <h2> 📌 What is DocuScopeAI? </h2> <p>DocuScopeAI is a web app that helps users (especially SMBs and freelancers) understand complex documents like NDAs, contracts, invoices, and terms &amp; conditions — without needing a lawyer.</p> <p>The app can:</p> <ul> <li>🧠 Extract raw text from PDFs and images (using OCR)</li> <li>✍️ Summarize documents with Hugging Face models</li> <li>❓ Answer questions about the content</li> <li>🌐 All through a simple, modern web interface</li> </ul> <h2> 🎯 Why I'm Building This </h2> <p>My main goals:</p> <ul> <li>Learn how to build modern full-stack apps</li> <li>Get good at FastAPI &amp; Hugging Face pipelines</li> <li>Dive into frontend dev (React, Tailwind, Next.js)</li> <li>Play with open-source LLMs</li> <li>Create something I’d actually use IRL</li> </ul> <h2> ⚙️ What I’ve Done So Far </h2> <p>✅ OCR-powered document ingestion (images + PDFs)<br> ✅ Summarization using facebook/bart-large-cnn<br> ✅ Raw text summarization endpoint<br> ✅ Question answering using OpenRouter’s LLaMA 3 model<br> ✅ Clean file structure with utils and separate routes<br> ✅ Started planning frontend (Next.js + Tailwind combo incoming)</p> <h2> 🔮 What’s Coming Next </h2> <ul> <li>✅ Connect frontend to the backend via REST</li> <li>📁 Add document versioning and history tracking</li> <li>🔐 Add user authentication</li> <li>🧩 Extract key clauses (like dates, names, legal terms)</li> <li>✨ Improve the UI with animations and polish</li> <li>📦 Deploy the app (thinking Railway or Fly.io for backend)</li> </ul> <h2> 🛠 Tech Stack </h2> <ul> <li>🧠 Hugging Face Transformers for AI tasks</li> <li>🧾 PyTesseract + pdf2image for OCR</li> <li>🚀 FastAPI for the backend</li> <li>💻 Next.js + Tailwind (soon) for the frontend</li> <li>🗃 PostgreSQL (coming later for storage)</li> <li>☁️ Hosting with Vercel + Railway (probably)</li> </ul> <p>Whole project is available on github: <a href="https://github.com/urewo396/DocuScopeAI" rel="noopener noreferrer">https://github.com/urewo396/DocuScopeAI</a> if you want to check it out or make some changes for yourself feel free to do it!</p> <p><strong>Thanks for reading!</strong><br> If you’ve built something similar or have ideas for cool features, hit me up in the comments 👇<br> And if you're also just vibing through learning FastAPI or frontend like me, let’s connect 💬</p> <p>Until next time — peace out 🧠⚡</p> programming python highschool fullstack