doc2mp3

Convert any document into an MP3 using Microsoft Edge neural text-to-speech.

Drop a PDF, Word doc, Markdown file, or plain text — get a high-quality MP3 back.

Quick Start

git clone <repo-url> doc2mp3
cd doc2mp3
npm install        # installs Node deps + sets up edge-tts automatically
npm start          # starts the web app at http://localhost:3456

That's it. Open http://localhost:3456 in your browser.

Requirements

• Node.js 18+ — nodejs.org
• Python 3.8+ — python.org (used for the TTS engine; auto-configured on npm install)

Both must be on your PATH. No API keys needed.

Supported Formats

Format	Extensions
PDF	.pdf
Word	.docx, .doc
Text	.txt, .csv, .log
Markdown	.md, .markdown
HTML	.html, .htm
Rich Text	.rtf
JSON	.json

Voices

9 neural voices are available in the UI:

Voice	Accent
Aria, Jenny, Ana	US Female
Guy, Christopher, Eric	US Male
Sonia	British Female
Ryan	British Male
Natasha	Australian Female

Speed is adjustable from 0.5x to 2.0x.

How It Works

1. Upload a document through the drag-and-drop web UI
2. The server extracts text using format-specific parsers (pdf-parse, mammoth, cheerio, marked)
3. Text is sent to Microsoft Edge's free neural TTS service via edge-tts
4. The resulting MP3 plays in-browser with a download button

Configuration

Env Variable	Default	Description
PORT	3456	Server port

Project Structure

doc2mp3/
├── server.js            # Express server
├── lib/
│   ├── parsers.js       # Document text extraction
│   └── tts.js           # Edge TTS integration
├── scripts/
│   └── setup-tts.js     # Auto-installs edge-tts on npm install
├── public/
│   └── index.html       # Web UI
├── uploads/             # Temp upload dir (auto-created, gitignored)
└── output/              # Generated MP3s (auto-created, gitignored)

License

MIT