textbook-to-note skill
A pipeline turns PDF and EPUB textbooks into markdown and writes Obsidian notes with book citations and figures.
- 111 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Python scripts read your books and write files into the notes vault
- Excerpt text reaches a cloud model when the agent writes a note
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/textbook-to-noteDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Python scripts read your books and write files into the notes vault
- Excerpt text reaches a cloud model when the agent writes a note
This is third-party code. Review the repository files before installing.
What it does
A local pipeline of five stages. The `converter/convert.py` converter extracts text from PDF and EPUB through PyMuPDF, detects pages with a corrupted text layer, restores column order and sends only problem pages to OCR. The figure module crops figures and passes them through a check. Two skills in `skills/`, `textbook-to-md` and `figure-remap`, and the algorithm in `workflows/note-writing.md` lead the agent from a book to a note where every claim is tied to a book and chapter. Heavy work spends no tokens, and the expensive model is used only for synthesis.
Who it is for. Medical students, researchers and teachers with dozens of reference PDFs who keep notes in Obsidian and want verifiable notes instead of a model's recall.
Good fit when
- Forty reference books cannot be read whole and the needed topic is spread across chapters
- A scan or PDF with broken fonts yields garbage instead of text
- A two-column textbook mixes up lines under ordinary extraction
- A note must cite a book and chapter for every claim
- Diagrams and tables from the book belong in the note, not a description of them
Not a fit when
- There are only a couple of books and ordinary reading with search is enough
- Someone else's protected material must be processed without copying rights
- Semantic search is needed out of the box: the indexer is not part of the repository
Example request
Convert this PDF to markdown, find the chapters on heart failure across three books and write a note citing book and chapter.Limitations
The repository contains no indexer: meaning-based search needs the separate vault-search project or a compatible CLI. Scanned books require OCR (Surya, ollama with a small vision model), which prefers an NVIDIA GPU or Apple Silicon. Tested on Windows 11 and macOS. Convert only your own books, as publishers keep copying rights. The last commit is from 3 November 2026.
How to disable. Delete the `textbook-to-md` and `figure-remap` folders from `~/.claude/skills/` and the cloned repository. Clean converted markdown files, the search index and notes in the vault separately.
FAQ
How do profiles A, B and C differ?
A only converts books to markdown for grep search. B adds agent-written notes with citations and figures. C connects meaning-based search through a separate indexer.
Is a GPU required?
Not for ordinary PDFs with a text layer: a CPU is enough. A GPU helps with OCR of scanned books.
Where do my books go?
Conversion, OCR and embeddings run locally. Data leaves the machine only through the model you ask to write the note.
How do I avoid wasted tokens?
`AGENTS.md` forbids the agent from reading a whole book. It finds pages in ready markdown through grep or the index.
Where has it been tested?
On Windows 11 and macOS, with Python 3.10 or newer. The code handles Windows quirks itself, such as cp950 decoding of subprocess output.
Discussion
No questions or reviews yet.
Related
SkillsTurns a lecture recording into slide-illustrated notes and an HTML page where video, transcript and summary stay in sync.
A skill that rebuilds PDFs, decks and scans into page-aligned Markdown by checking extracted text against page images.
A Firecrawl skill and CLI that lets your agent convert Word, PowerPoint, Excel, OpenDocument, RTF, EPUB, CSV and PDF to Markdown on your own machine
doc-cleaner converts PDF, Office, Keynote, Numbers and EPUB into clean Markdown with intact tables and Big5 and CP950 encodings.
BookTrans translates a whole book in one run: scouting, context-aware translation, editing, footnotes and building to epub or fb2.
A skill and CLI that turn Markdown into DOCX, PDF, PPTX, XLSX, PNG and about a dozen more formats, all computed locally
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| textbook-to-note (this card) | no | 111 | ||
| Obsidian Knowledge Agent | no | 206 | files PDFs, slides and links into notes; here a textbook converter with column, OCR and figure checks | |
| pdf-mcp | no | not required | 171 | an MCP server searches and reads needed PDF pages on request; here a standalone pipeline with ready markdown and cited notes |
README in short
The project follows three rules: heavy work without tokens, strict checks instead of eyeballing and a book reference for every claim. The author has kept medical notes since med school.