SecureDocAI - AI-Powered Document Management System

• 

  01 Landing Page

• 

  02 Signup Page_Demo Page

• 

• 

• 

• 

• 

• 

• 

• 

• 

Inspiration

• Sending identical PDFs to every recipient wastes storage and clutters devices.
  
• Many people—especially seniors—prefer a concise summary over reading a full document.
  
• A single, up-to-date copy in the cloud plus on-demand AI summaries meets both needs.

What it does

• Lets each user upload up to 5 PDF files per day.
  
• Stores one master copy in Supabase Storage and writes metadata to Postgres.
  
• Runs an Edge Function that calls an LLM to create a short, plain-language summary.
  
• Displays file status and the generated summary in a React dashboard.
  
• UI text is localised in English, Arabic, French, and Spanish.

How we built it

Challenges we ran into

• Storage trigger missed auth.uid(), so no metadata rows were created after uploads.
  
• HTTP 406 on app_config—fixed by adding Accept: application/json to the fetch.
  
• Infinite loading spinner from route-level Suspense; solved with an explicit auth guard.
  
• “Empty offerings” in RevenueCat until products were activated and linked to the correct bundle.

Accomplishments that we're proud of

• End-to-end PDF → AI summary flow running on Supabase’s free tier.
  
• Subscription screens fully translated into four languages.
  
• Secure password-reset flow implemented by parsing the Supabase hash correctly.

What we learned

• Test storage triggers with real uploads; mocks can hide auth-context issues.
  
• API errors like 406 often come down to a missing header.
  
• Translation keys grow fast—define a clear naming scheme from day one.

What's next for SecureDocAI – AI-Powered Document Management System

• Batch upload with email alerts when summaries are ready.
  
• Team workspaces to share daily upload quotas.
  
• Audit log export for compliance reviews.
  
• Summary styles (executive vs. technical) selectable per document.

Built With

• react
• supabase
• tailwind
• vite