Open-source licences
The open-source software, models, fonts and word lists that run pdftomd in your browser, and their licences.
Last updated: 2026-11-13
pdftomd is built on the work of open-source projects. The converter, the OCR and the token counter all run in your browser using the components below. Each is used under its own licence, and nothing in our terms limits the rights those licences give you.
The full licence text of every component that reaches your browser is in third-party-licenses.txt.
Conversion and OCR
| Component | What it does here | Licence |
|---|---|---|
pdf-inspector (@firecrawl/pdf-inspector-wasm) by Firecrawl | Reads the PDF's text layer, layout and tables | MIT |
PDF.js (pdfjs-dist) by Mozilla | Draws pages for OCR and the page previews, and pulls out embedded images for the .zip | Apache 2.0 |
| ppu-paddle-ocr by PT. Perkasa Pilar Utama | Runs text detection and recognition | MIT |
| PaddleOCR models by the PaddlePaddle authors | The text recognition models (PP-OCRv6 tiny, PP-OCRv5 mobile, Korean PP-OCRv5), in ONNX form | Apache 2.0 |
ONNX Runtime Web (onnxruntime-web) by Microsoft | Runs the models in WebAssembly | MIT |
| gpt-tokenizer by Bazyli Brzoska | Counts tokens with the o200k_base encoding | MIT |
| fflate by Arjun Barrett | Builds the .zip downloads | MIT |
| markdown-it by Vitaly Puzrin and Alex Kocharin | Renders the Markdown preview | MIT |
Word list
The converter rejoins words that a PDF splits across lines or spaces out letter by letter, and checks each repair against the words the document itself uses and about 10,700 common English words from SCOWL (sizes 10 and 20). SCOWL is Copyright 2000-2016 Kevin Atkinson and is used under its permissive licence; its full notice is at the end of third-party-licenses.txt.
Interface and fonts
| Component | Licence |
|---|---|
| React | MIT |
| TanStack Router, Start and Query | MIT |
| Base UI | MIT |
| Paraglide JS | MIT |
| Tailwind CSS | MIT |
| Lucide icons | ISC |
| Public Sans, Literata and JetBrains Mono fonts | SIL Open Font License 1.1 |
The text file lists the smaller libraries these depend on, with their licences.
Sample PDFs
The three sample files on the converter are excerpts from works of the United States government, which are not subject to copyright in the United States:
- Artificial Intelligence Risk Management Framework (AI RMF 1.0), NIST, 2023
- Apollo 11 Mission Report, NASA, 1969, a scan of the typewritten report
- Publication 15-T, Federal Income Tax Withholding Methods, Internal Revenue Service
Their names appear only to describe the documents; the agencies have no connection with this site.
Names of other tools
MarkItDown is a project of Microsoft, PyMuPDF and PyMuPDF4LLM are developed by Artifex, Docling is hosted by the LF AI & Data Foundation and Marker is made by Datalab. Obsidian and every other product named on this site belong to their owners. We use these names only to describe and compare the tools, and pdftomd isn't affiliated with any of them.
Questions
If you think we have left out a notice or credited something wrongly, write to hello@pdftomd.net and we'll fix it.