FillWizard

AI-powered PDF form autofill SaaS built from scratch with Next.js and TypeScript, from the marketing site to the authenticated dashboard.

Company: Personal project
Role: Frontend Lead
Stack: Next.js TypeScript Tailwind CSS ShadCN UI NestJS Prisma Redis/BullMQ
Highlights: Built the full frontend from scratch, marketing site, auth flows, and dashboard, in Next.js App Router and TypeScript.
Shipped a multilingual UI (English, French, Arabic) with full RTL support for Arabic.
Built the PDF review and field-mapping UI, letting users confirm autofill suggestions before export.
Integrated the frontend with a NestJS/BullMQ backend for async PDF processing and OCR jobs, with clear pending/failure states in the UI.

FillWizard landing page, PDF form autofill product

Live preview

What it does

FillWizard is a PDF form autofill product: upload a PDF, detect its form fields (AcroForm or OCR for scanned/flat PDFs), map them to a saved identity profile, review the suggested values, and export a filled PDF. Built for repetitive form-heavy workflows: visa packets, school forms, ACORD insurance claims, and government paperwork.

I built this from scratch as Frontend Lead, on top of a Next.js and NestJS monorepo I also set up.

Frontend

Next.js (App Router) + TypeScript for the marketing site, auth, and dashboard.
Tailwind CSS + ShadCN UI for a consistent, accessible component system.
i18next, with English, French, and Arabic, including full RTL layout support for Arabic.
Built the core product flows: PDF upload, field detection review, profile-to-field mapping, and export, with loading, retry, and error states for each async step.

Working across the stack

Wired the frontend to a NestJS API with Prisma/PostgreSQL and a BullMQ/Redis job queue, so the UI reflects real background-job state (queued, processing, done, failed) instead of guessing.
Used MinIO/S3-compatible storage for uploaded PDFs and generated exports, with signed URLs handled through the API.