v0.2.2: OCR Translation

Highlights

🖼️ Image OCR

• Right-click "Extract Text (Flowers)" on any image to extract text via LLM Vision
• Extracted text flows into the popover for translate, polish, notes, and ask
• Customizable ocr_system / ocr_user prompts in Settings
• Requires a Vision model (e.g. GPT-4o, Claude 3, Gemini 3 Flash)

📄 PDF Reader Enhancements

• Toolbar: Download, print, search, fullscreen, zoom, dark mode, page jump
• GitHub/GitLab fix: Auto-convert blob URLs to raw for embedded PDFs
• Performance: IntersectionObserver lazy loading for long documents
• Stability: Fixed "Canvas in use" conflicts and scaling crashes
• UX: Improved loading and error states

📺 Demo Video

• Added demo video showcasing the full Flowers workflow

Technical

• Bootstrap loads prompt overrides before first request
• Star History chart in README footer