Marker is an open-source tool from Datalab that converts PDFs and other documents into Markdown, JSON or HTML, keeping tables, equations and code readable for AI apps.
In plain words
Marker reads PDFs, images, PowerPoint, Word, Excel, HTML and EPUB files in any language and rewrites them as clean Markdown, JSON, chunks or HTML. It formats tables, forms, equations and code blocks, saves images, and strips page clutter such as headers and footers. An optional hybrid mode adds a language model to fix harder cases, such as tables that run across pages. It runs on a GPU, a CPU or Apple silicon.
A quick example
Imagine a maths textbook PDF you want to use in a study chatbot. Marker converts it into Markdown with the equations written out as maths and the headers and page numbers removed, so the text can be chunked and searched cleanly.
Where you'll meet it
- Converting PDFs for RAG and training data.
- Extracting tables, equations and forms.
Related
Sources
- Docsdatalab-to/marker, Datalab (GitHub) · read 21 Oct 2026
- DocsDatalab Marker, Datalab · read 21 Oct 2026
Checked by WhatIsAI against the sources above · 21 Oct 2026. Spotted an error? Report it.