Last updated: 2026-06-10
Privacy Policy
How MarkItDown AI handles temporary uploads, analytics, privacy boundaries, and non-affiliation notices.
File processing
Uploaded files and pasted content are used only to perform the requested document-to-Markdown conversion.
V2.0 does not use persistent file storage. Files should be written to an isolated temporary directory by the conversion service and cleaned up after the request completes or fails.
No AI training on files
MarkItDown AI does not train models on uploaded files or converted Markdown output in this release.
The default conversion flow does not send file contents to an LLM provider.
Analytics boundary
Analytics may collect page route, file extension, size bucket, duration bucket, and error code to improve the product.
Analytics must not collect file contents, converted Markdown, full filenames, or private text snippets from documents.
Security limits
V2.0 uses a default whitelist for PDF, DOCX, PPTX, XLSX, CSV, JSON, XML, HTML, HTM, TXT, and MD files.
OCR, ZIP archives, audio transcription, video, YouTube URLs, remote URL fetching, and hosted MCP server features are disabled by default or reserved for future versions.
Non-affiliation
MarkItDown AI is an independent tool site and is not affiliated with Microsoft. MarkItDown is an open-source project associated with Microsoft. This site provides a browser-based conversion workflow built for AI-ready Markdown use cases.
Contact
For privacy, file handling, or support questions, contact support@markitdownai.org.