📊 How Screenshot OCR Works
1. Upload screenshot (Twitter, article, anything with text)
2. Tesseract.js extracts text in your browser
3. POST to your local Ollama server
4. AI extracts structured insights
5. Idea appears in your archive
Privacy: OCR runs 100% in your browser - images never leave your device until you click "Extract Insights"
Tesseract.js (OCR) → Local Ollama (AI extraction) → GitHub Pages (publishing)