Handwriting-Recognition

---

Dataview:

list from [[]] and !outgoing([[]])

---

🧲 Published

Deployment:

• https://handwriting-wrapper.vercel.app

GitHub:

• https://github.com/diegotyner/handwriting-wrapper

🧾 Project Description

Blurt

I wanted a little text extractor for when its too annoying to highlight PDF text.

I might want to extend it to actually extract text/notes!

• Send images to a flagship LLM
  • e
• Maybe use the extracted text to make some sort of knowledge management application?
  • Notion database?
  • Obsidian?
  • Class folders?

1. User Uploads Handwritten Note
    - Mobile/web app captures image → uploads to Firebase Storage or directly to Notion.
2. OCR Processing
    - Use Google Cloud Vision (free tier) or Tesseract (free, less accurate) to extract text.
3. LLM Structuring
    - Send text to GPT-3.5 to:
        - Generate a title, summary, and tags.
        - Identify related notes (for backlinks).        
4. Save to Notion/Firebase
    - Create a Notion page or Firestore document with structured data.    
5. Update Vector DB
    - Generate embeddings for the note → store in Pinecone/Chroma.    
6. Semantic Search
    - Users query notes via text → search vector DB for matches

Maybe I could even tie it into Obsidian-LLM-Integration?

Features

• Copy / paste or import image files
• Copy paste extracted text out

Official

🎯 Objective

📂 Project Logs

🎟 Features

Existing

Todo

🔗 -> Links

Resources

• Put useful links here

Connections

• Link all related words