Yoduhu

• 

  Yodohu – Your AI-powered document hub. Smart organization and instant access to all your files, powered by AI.

• 

  Find any document in seconds. Yodohu’s Smart Search understands intent and finds what you need—instantly.

• 

  Upload your documents effortlessly. Just drag & drop. Yodohu supports instant PDF processing—no setup needed.

• 

  Dark mode, same power. A polished, responsive UI for managing, searching, and tagging your documents.

Inspiration

I was at the buildersklub event in Berlin last week, and one discussion really stuck with me: the potential of automation and AI is still not accessible to many people - mostly because their data isn’t digitized or their databases aren’t properly maintained. Especially in Germany, where I’m from, a lot of important documents are still handed out on paper.

That made me think — wouldn’t it be great if all of this could be automated? Imagine a system that automatically organizes and maintains all your documents. No folders, no effort. And whenever you need something, it’s always accessible - no matter where you are.

What it does

Yodohu is your personal document hub powered by AI. You can upload PDFs (and soon Images), and Yodohu will:

• Automatically extract content and metadata
• Categorize the document intelligently
• Generate relevant tags and a short summary
• Make everything searchable via semantic search & smart ai search
• Let you mark important documents as favorites

How I built it

I used

• Bolt.new as the main development environment
• Supabase for database, auth, and storage
• OpenAI GPT-3.5 for text analysis, summarization, tagging & categorization
• TailwindCSS for styling and theming
• Lucide Icons for modern UI consistency
• Netlify for deployment

Challenges we ran into

• Handling inaccurate AI classification and tweaking prompt logic
• Managing file types and formats during upload validation
• Integrating Supabase securely while maintaining real-time UX
• Keeping the UI minimal and intuitive while packing in features

Accomplishments that we're proud of

• Fully working MVP with advanced AI document parsing within 2 days (around 10h)
• Intuitive UX that feels like a real product, not a prototype
• Smart Search feature which let's you prompt what you searching for in any language
• Landing page, dark mode, and real-time uploads — all polished and styled

What we learned

• How easy it is to use Bolt.new effectively as a full-stack platform and build an actual working product
• How to write more effective prompts for AI extraction & classification
• Deep understanding of integrating Supabase auth/storage (first time for me)
• Optimizing UX for document-heavy workflows

What's next for Yoduhu

• OCR support for image-based documents
• Smart reminders for due dates & expiring documents
• User-defined custom tags & categories
• Secure sharing and collaboration features
• Mobile app for the on-the-go access
• Integrations like calendar, reminder, sharing, automatic imports from 3rd parties

I'm really shocked by how much of a product I could build in such a little time with the help of bolt.new and ChatGPT. I'm truly impressed and got a new understanding of vibe coding and its possibilities. I'm super happy to share this product with you and looking forward to the feedback.

Built With

• bolt.new
• openai
• react
• supabase
• tailwind
• typescript