Neurocloud156 downloadsConvert handwritten PDF notes into formatted Markdown with Gemini document understanding.
Turn handwritten PDF notes into clean Obsidian notes without doing the copy-paste dance.
Handwriting PDF is a small Obsidian plugin for those PDFs that are full of useful handwritten notes but are annoying to search, skim, or reuse. Pick a PDF, run the plugin, and it asks Gemini to read the handwriting, clean up the obvious typos and punctuation, preserve the structure it can understand, and drop everything into a fresh Markdown note.
Current version: 0.1.20.
This project was built in collaboration with AI using OpenAI Codex.
By default, Handwriting PDF keeps the workflow simple:
YYYY-MM-DD - NoteTitle.md.YYYY-MM-DD - NoteTitle as the default.#Todo by default, using a configurable tag.Handwriting PDF Notes unless you choose a different folder.The goal is simple: handwritten PDF in, useful Obsidian note out.
Install the plugin in your vault, open the Handwriting PDF settings, and add your Gemini API key. The default model is gemini-3.1-flash-lite. You can also choose another PDF-friendly handwriting model from the dropdown.
That is the main setup. After that, select a PDF and run Create handwriting note.
Generated notes are meant to be easy to scan:
# 2026-06-07 - Example Title
## Details
- Source PDF: [[example.pdf]]
## Summary
...
## Transcription
...
## Source PDF
[[example.pdf]]
If you turn on Embed PDF in note, the final PDF section uses an embed instead:
![[example.pdf]]
There is also an OCR-enhanced PDF feature. It is on by default.
When enabled, the plugin creates a separate PDF copy with an invisible searchable text layer. The original PDF is not changed. By default, the Markdown note is created first and the OCR-enhanced PDF is created afterward in the background.
To keep things fast, the plugin checks whether the PDF already has a text layer:
This plugin needs your own Gemini API key. When you run a conversion, the selected PDF is sent to Google's Gemini API for handwriting recognition, cleanup, formatting, and optional summary generation.
The API key is saved in Obsidian's local plugin data for the current vault. The plugin does not include a bundled API key, does not run ads, and does not collect telemetry.
Handwriting recognition is only as good as the source PDF and the model response. The plugin tries to preserve structure, but messy handwriting, cramped margins, and complex layouts can still need a quick human pass.
This plugin bundles pdf-lib under the MIT license; see PDF-LIB-LICENSE.md.