Dev.to
7/12/2026
Annotation extractor tool uses Gemini to turn PDF/DOCX highlights into themed study notes
Original: I Read Too Much and Forgot Everything So I Built a Tool That Turns My Annotations Into Something I'll Actually Reread.
Short summary
A developer built a tool that extracts highlights, underlines, strikeouts, and comments from PDF and DOCX files into a structured list organized by color and type. It uses Gemini Flash-Lite to cluster highlights into themes and generate synthesized study notes, then exports to Markdown, Word, PDF, or Anki flashcards. The architecture uses a shared data model with format-specific extractors and a FastAPI backend serving both API and frontend.
- •Extracts annotations from PDF/DOCX into a unified color-coded structured list
- •Gemini Flash-Lite clusters highlights into themes and writes connected summary paragraphs
- •Exports to Markdown, Word, PDF, and Anki flashcards
Generated with AI, which can make mistakes.
Is this a good recommendation for you?
