Skip to content
Convert tools

File to Markdown — PDF, Word, Excel to MD for LLMs, No Upload

Convert PDF, Word (DOCX), PowerPoint, Excel, HTML, CSV, JSON or EPUB to Markdown for ChatGPT and other LLMs. Free, runs Microsoft MarkItDown in your browser, no upload.

Free · No sign-up · Files are converted in your browser and never uploaded

Turn PDF, Word, PowerPoint, Excel and other documents into clean Markdown you can paste into ChatGPT, Claude or a notes app. Microsoft MarkItDown runs in your browser, so your files never leave this device.

or drag and drop files here

PDF, DOCX, PPTX, XLSX, XLS, HTML, CSV, JSON, XML, EPUB, IPYNB, ZIP and more · Up to 20 files, 50 MB each

TXT, Markdown, JSON and XML are processed directly. Other formats download Python and conversion packages on first use; download size depends on the format. File contents are not uploaded.

Your Markdown will appear here

Add one or more files. Headings, lists, tables and links are kept wherever the source has them.

  • PDF
  • DOCX
  • PPTX
  • XLSX
  • XLS
  • HTML
  • CSV
  • JSON
  • XML
  • RSS
  • EPUB
  • IPYNB
  • ZIP
  • TXT
  • MD

Check the result before relying on it. Page layout, images and charts are not converted, and scanned pages have no text to extract; use PDF OCR for scans.

How to use

  1. Choose or drop one or more files: PDF, Word (DOCX), PowerPoint (PPTX), Excel (XLSX or XLS), HTML, CSV, JSON, XML, EPUB, Jupyter notebooks or a ZIP of these files.
  2. Wait while the file is converted. The first run loads the Python runtime and MarkItDown; later files and visits are faster.
  3. Review the Markdown, switch to Preview to see it rendered, then copy it or download a .md file. With several files you can also download one combined .md.

Example

report.docx (Heading 1 “Quarterly report”, “Revenue grew 12%” with 12% in bold, two bullet points)
→
# Quarterly report

Revenue grew **12%**

* First item
* Second item

sales.xlsx (sheet “Sales”)
→
## Sales
| Region | Sales |
| --- | --- |
| EU | 42 |
| US | 58 |

Details and limitations

What it does: File to Markdown converts documents into Markdown, a plain-text format that keeps structure such as headings, lists, tables and links. Markdown is compact and easy for ChatGPT, Claude, Gemini and other large language models (LLMs) to read, and it works in RAG pipelines, notes apps and documentation sites.

Supported formats: PDF (text layer and simple tables), Word .docx, PowerPoint .pptx (slide text, tables and speaker notes), Excel .xlsx and .xls (one table per sheet), HTML, CSV, JSON, XML, RSS and Atom feeds, EPUB, Jupyter .ipynb, plain text and ZIP archives of these files. Legacy .doc and .ppt files, images, audio and Outlook .msg are not supported in the browser.

Privacy: conversion runs locally with Microsoft MarkItDown (MIT license) inside Pyodide, a WebAssembly build of Python, in a Web Worker. The page downloads code (the runtime from cdn.jsdelivr.net and Python packages from pypi.org), but your files and the Markdown are not uploaded. Site analytics and advertising make their own requests.

TXT, Markdown, JSON and XML are processed directly. Other formats download Python and conversion packages on first use; download size depends on the format. File contents are not uploaded.

Limits: up to 20 files of 50 MB each per batch; very large files can exhaust memory on phones. Page layout, fonts, images and charts are not reproduced. Scanned PDFs have no text layer, so run PDF OCR first. PDF tables are detected heuristically and may need cleanup.

FAQ

How do I convert a PDF to Markdown?

Drop the PDF on this page. Its text layer is extracted as Markdown, and simple tables become Markdown tables. Copy the result or download it as a .md file. Scanned PDFs need OCR first.

How do I convert a Word document to Markdown?

Choose the .docx file. Headings, bold and italic text, lists, tables and links become Markdown. Old .doc files are not supported; save them as .docx in Word or LibreOffice first.

Why convert files to Markdown for ChatGPT or other LLMs?

Markdown keeps the document structure in plain text, so a language model can tell headings, lists and table cells apart while using fewer tokens than raw HTML. Paste the result into a chat or add the .md file to a knowledge base.

Are my files uploaded?

No. The converter runs in your browser in a Web Worker. Only the converter code is downloaded; file contents and results stay on your device.

What is MarkItDown?

MarkItDown is an open-source Python library from Microsoft that converts office documents and other files to Markdown for LLM use. This page runs MarkItDown 0.1.8 in Pyodide, so no server or installation is needed.

Why is the first conversion slow?

TXT, Markdown, JSON and XML are processed directly. Other formats download Python and conversion packages on first use; download size depends on the format. File contents are not uploaded.

Can it convert images or scanned documents?

No. Images need OCR or an AI captioning service, which this browser version does not include. Use Image to Text or PDF OCR to extract text from scans first.

Format references

Tool page updated: