Drop a document here
or browse your files — converted entirely in your browser, nothing is uploaded
Source
—
Your document preview will appear here.
Markdown — editable
Supported formats
Document → Markdown
| Format | Engine | Output quality | Notes |
|---|---|---|---|
.docx |
mammoth.js → turndown | Excellent | Headings, lists, bold/italic, tables. The strongest path — often cleaner than markitdown because mammoth reads the docx structure natively. |
.txt |
passed through | Exact | No transformation; plain text is already Markdown-compatible. |
.csv |
SheetJS | Good | Becomes a GFM pipe table. |
.xlsx / .xls |
SheetJS | Good | One Markdown table per sheet; multi-sheet workbooks get a ## heading per sheet. |
Markdown → Document
| Direction | Engine | Output quality | Notes |
|---|---|---|---|
Markdown → .docx |
marked → docx | Good | Switch to “Markdown → Document”, then type, paste or drop a .md file. Headings, bold/italic, lists, tables, links and blockquotes carry across to a Word .docx. Very advanced Markdown or embedded HTML may render simply. |
Not supported by design
| Format | Why | What to use instead |
|---|---|---|
.pdf |
A PDF stores positioned glyphs, not structure. Rule-based extraction flattens headings, lists and tables into a wall of text. | Drop it into an AI assistant (Claude, ChatGPT) and ask for “clean Markdown, preserving heading hierarchy and tables” — current models do this far better than rule-based extraction. For bulk or automated work, Microsoft’s markitdown CLI is the server-side option. |
.doc |
Legacy binary format. mammoth.js only handles the modern XML-based .docx. |
Open in Word and re-save as .docx, then drop it here. |
.pptx |
Out of scope for now. | — |
Why pure client-side? No server means no upload endpoint, no auth, no data storage, nothing to secure. Files never leave the browser — the right model for potentially sensitive client documents.