datalab-to/marker

Health v2 · maintenance only

Adoption not part of the score

40358 stars · 2918 forks observed · 2026-10-11

What it is AI-extracted, prompt v1, taxonomy v1, 2026-09-28, confidence 0.97

Marker is an open‑source Python tool that converts PDFs and other document formats into markdown, JSON, HTML, and structured chunks. It supports tables, equations, images, and can optionally boost accuracy with LLM prompts.

Use cases

• Convert PDF files to markdown for documentation pipelines
• Extract structured JSON data from PDFs, Word docs, and spreadsheets
• Batch‑process large collections of PDFs with high accuracy
• Preserve tables, forms, and equations when converting documents
• Boost OCR accuracy using a custom LLM prompt
• Generate HTML previews of scanned documents

When to choose

• You need high‑precision conversion of complex documents (tables, math, multi‑column layouts)
• You want output in markdown or JSON for downstream processing
• You require multilingual support across 90+ languages
• You prefer an open‑source solution that can run on CPU, GPU, or MPS
• You want the option to enhance results with your own LLM

When to avoid

• You need a graphical UI with no command‑line interaction
• Real‑time conversion on low‑power edge devices is required
• Only plain text extraction without formatting is needed
• You require built‑in document editing capabilities

Facets

cli-tool · maturity active

parser ocr llm-inference file-upload image-processing artificial-intelligence data-science python windows cli pdf-to-markdown markdown-converter json-export llm-boost high-accuracy multilingual document-intelligence linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="datalab-to/marker")