MarkItDown.tech
Free online converter that turns PDF, Word, PowerPoint, Excel, HTML, EPUB, URLs and images into clean, AI-ready Markdown, with private in-browser processing and AI OCR for scanned PDFs.
Overview
MarkItDown.tech is a browser-based document converter that turns PDFs, Office files, HTML, EPUB, URLs, and images into clean Markdown for AI workflows. It's built on top of Microsoft's open-source MarkItDown Python library, but packaged as a no-install web app rather than a CLI or SDK — a third-party hosted interface, not an official Microsoft product. Files are processed as temporary files and deleted after conversion, with optional local (in-browser) processing for supported formats, and both standard and AI-enhanced OCR for scanned or image-embedded documents.
The Verdict
Who Should Use MarkItDown.tech?
Best For
- One-off conversions of PDFs, Office docs, or URLs into Markdown for RAG or LLM prompts
- Non-developers who want MarkItDown's output without installing Python
- Quick scanned-document cleanup via built-in OCR
Not Ideal For
- Batch or pipeline conversion at scale (free tier caps at 3 files/batch, 20MB each)
- Teams needing an API or self-hosted deployment — use the underlying open-source library directly
- Highest-fidelity table/layout extraction — not benchmarked as strongly as dedicated parsers
What's Great
- No installation, no Python, no account needed to start converting
- Covers a wide input range: PDF, DOCX, PPTX, XLS/XLSX, HTML, EPUB, CSV/TSV, JSON, XML, images, and public URLs
- Files deleted after conversion; optional local in-browser processing for privacy-sensitive formats
- AI-enhanced OCR available for scanned PDFs and image-embedded text
Watch Out For
- Free tier limited to 3 files per batch, 20MB per file, and 10 OCR pages/month
- Independent benchmarking of the underlying MarkItDown library scored it 57/100 on PDF conversion quality, behind dedicated parsers like MinerU and Marker (86/100)
- Not an official Microsoft product — it's a third-party wrapper around the open-source library
- Minimal public review coverage (G2 listing has 0 ratings)
Pricing
View all features & details
Supported Input Formats
- PDF, DOCX, PPTX, XLS/XLSX
- HTML/HTM, EPUB
- TXT/MD, CSV/TSV, JSON, XML, IPYNB
- PNG, JPG, WebP images
- Public URLs
Privacy
- Files processed as temporary files, deleted after conversion
- Source files not stored or logged
- Optional local (in-browser) processing for supported formats
Underlying Technology
- Built on Microsoft's open-source MarkItDown library (MIT licensed, one of the most-starred document-conversion projects on GitHub)
- Not operated or maintained by Microsoft
How It Compares
| Feature | MarkItDown.tech | Docling | Unstructured |
|---|---|---|---|
| Install Required | No (browser-based) | Yes (Python) | Yes (Python) |
| Self-hosted | No | Yes | Yes (Business) |
| Free File Limit | 20MB | No limit | Varies |
| OCR | Yes (AI-enhanced) | Yes | Yes |
| PDF Conversion Quality | 57/100 (underlying lib) | Best-in-class tables | High |
| Pricing | Freemium | Free (MIT) | Freemium |