Document OCR to Markdown converts your uploaded txt, csv, docx, xlsx, pptx, odt, ods, odp, or pdf documents into one editable Markdown file (named YYYYMMDDHHMMSS.md). It runs as a multi-round AI Agent workflow: each new conversation creates a notebook workspace named YYYYMMDDHHMMSS in the background, the system first renders every page of the document into a page image, and the site's configured vision model then performs optical character recognition on each page in order. Pictures embedded in the document are cropped into separate image files, linked in the Markdown, and captioned with the text found inside each figure; tables are re-created as properly formatted Markdown tables; heading levels and reading order are preserved as much as possible. When all pages are assembled in order, the .md file is generated with a download link, and each cropped figure gets its own download link as well. Usage is simple: upload one or more documents and press send. The conversion streams each round's progress and tool actions in real time, can be interrupted at any time with the stop button, and you can continue the same conversation to request adjustments. The Markdown content keeps the source document's original language — it is a conversion, not a translation. You can also open this conversation's notebook on the Notebook page to review the page images and generated files in its workspace. Document OCR to Markdown runs many rounds of actions in a row and calls the vision model page by page, using a large amount of tokens; at least 100,000 credits are required in your account to use this feature. Maximum total upload size per request: 256MB (all files and message content combined). Output content formats: Text/Image Accepted document formats: txt, csv, docx, xlsx, pptx, odt, ods, odp, pdf. File output: an editable Markdown file named YYYYMMDDHHMMSS.md (cropped figures are saved as separate image files with download links).