Pourdown

Turn any document into clean, editable Markdown.

A desktop Markdown editor that imports Word files, spreadsheets, PDFs, and presentations in one click, then lets you write and edit with a live visual preview — free, offline, and open source.

🌐 Website

Demo

demo.mp4

Importing a PDF and getting editable Markdown back in one click.

Screenshots

Import from Other Formats

Most writing lives in Word documents, spreadsheets, or slide decks — formats that are hard to version-control, collaborate on, or publish as-is. Pourdown lets you import any of them directly into an editable Markdown document, so you can clean up, restructure, and export without manual copy-pasting.

Why Markdown Import?

Files are converted to Markdown before processing to minimize token usage. Community benchmarks show Markdown is ~15% more token-efficient than JSON, and up to 96% more efficient than raw PDF when fed to LLMs

This approach was inspired by Microsoft's MarkItDown; see markdown-import.md for how Pourdown's Rust implementation works and how it differs.

What gets converted

Format	What's preserved	Known limitations
Word (.docx)	Headings (styles + outline level), bold / italic / strikethrough, nested bullet and numbered lists, tables, hyperlinks, embedded images	Vector images (EMF/WMF) can't be displayed; tracked changes and comments are dropped; a TOC placeholder is inserted
Spreadsheet (.xlsx / .xls / .ods)	Each sheet becomes a section with a full GFM table; date columns are auto-detected and formatted as ISO dates; embedded images are extracted	Capped at 500 rows per sheet; images can't be mapped to a specific cell/sheet
PDF	Headings inferred from font-size ratios; paragraph flow sorted top-to-bottom; tables are detected via column-alignment clustering and ruling-line boxes and rendered as GFM tables; Table of Contents entries render as a bulleted list; equations are demarcated as $…$ / $$…$$; embedded images are extracted with adjacent figure/table captions used as alt text	Text-based PDFs only — scanned / image PDFs are not supported; complex multi-column layouts may reorder; tables with deeply nested/bulleted wrapped cells may fall back to plain text; math is Unicode symbols demarcated in $…$/$$…$$, not reconstructed LaTeX
PowerPoint (.pptx)	Slide titles become # headings; body text becomes paragraphs, one slide per section; embedded images are extracted	Animations are not captured; vector images (EMF/WMF) can't be displayed

Extracted images are saved as sidecar files next to the imported document (an assets/ folder, relocated to <name>.assets/ alongside the .md once saved) and render live in the editor.

Import converts content to Markdown. It is not a round-trip format converter — exporting back to the original format will not restore the original layout exactly.

Features

• Document Import — Convert Word, Excel, PDF, and PowerPoint files to Markdown in one click
• Visual Markdown Editing — Write and edit Markdown the way you see it, without dealing with raw symbols
• Source Mode — Toggle to raw Markdown text at any time
• File Management — Open, save, and manage your Markdown files using native system dialogs
• Rich Text Formatting — Bold, italic, lists, code blocks, blockquotes, and more
• Auto-save — Your work is saved automatically at regular intervals
• Find & Replace — In-document search with replace support; cross-file search in the sidebar
• Multiple Themes — Choose from seven built-in UI themes
• i18n — English and Traditional Chinese interface

Tech Stack

• Frontend: React 18 + TypeScript
• Editor: Tiptap v2 with extensions
• Desktop: Tauri v2
• State Management: Zustand
• Styling: Tailwind CSS
• UI Components: Custom components with shadcn/ui patterns
• Build Tool: Vite

Getting Started

Prerequisites

• Node.js (v20 or later)
• pnpm (or npm/yarn)
• Rust (for Tauri desktop builds)

Development

Start the development server with hot reload:

pnpm dev

For Tauri-specific development:

pnpm tauri dev

Building

Create an optimized production build:

pnpm build

Create the desktop application:

pnpm tauri build

Contributing

Contributions are welcome! See CONTRIBUTING.md for setup steps, the test-first workflow for conversion changes, and the PR process. The docs site has a full user guide (English / 繁體中文) if you want to understand a feature in more depth than this README covers.

Acknowledgements

The Markdown Import feature was inspired by Microsoft's MarkItDown — both projects are MIT-licensed. Pourdown is an independent reimplementation in Rust (not a fork or port); see markdown-import.md for details on how the two differ.

License

MIT