@mdgate/rtf

Convert RTF to Markdown in TypeScript.

@mdgate/rtf reads Rich Text Format files directly in JavaScript and converts them into GitHub-Flavored Markdown, without Python, native addons, WASM, or Word.

Works in Node.js, Cloudflare Workers, Edge runtimes, and browsers.

npm install @mdgate/rtf

import { toMarkdown } from '@mdgate/rtf';

const markdown = await toMarkdown(bytes);

bytes is a Uint8Array, so the document can come from a file upload, object storage, an HTTP request, a browser file picker, or anywhere else your application gets bytes.

Why @mdgate/rtf

RTF still shows up in email, older Word "Save As" output, and files that wear a .doc extension but start with {\rtf.

@mdgate/rtf is an RTF reader written for the same runtime as your application:

• Pure TypeScript
• RTF → Markdown locally
• No Python runtime
• No native addons
• No WASM runtime
• Zero third-party runtime dependencies
• Works with raw Uint8Array input
• Detects RTF from the file contents, not only the filename

What it extracts

@mdgate/rtf lexes RTF control words and rebuilds a shared document model.

The converter handles RTF-specific concerns including:

• paragraphs and headings
• bold, italic, and strikethrough
• lists
• tables
• Unicode and legacy code-page text
• RTF content stored under a .doc name

The output is Markdown that can be searched, indexed, chunked, cached, or passed directly to an AI agent.

Node.js

import { readFile } from 'node:fs/promises';
import { toMarkdown } from '@mdgate/rtf';

const bytes = new Uint8Array(await readFile('notes.rtf'));
const markdown = await toMarkdown(bytes);

console.log(markdown);

Browser

import { toMarkdown } from '@mdgate/rtf';

const file = input.files![0];
const bytes = new Uint8Array(await file.arrayBuffer());

const markdown = await toMarkdown(bytes);

Cloudflare Workers and Edge runtimes

import { toMarkdown } from '@mdgate/rtf';

export default {
  async fetch(request: Request) {
    const bytes = new Uint8Array(await request.arrayBuffer());
    const markdown = await toMarkdown(bytes);

    return new Response(markdown, {
      headers: {
        'content-type': 'text/markdown; charset=utf-8',
      },
    });
  },
};

Format detection

You do not need to trust the file extension.

@mdgate/rtf recognizes RTF from the opening {\rtf group.

const markdown = await toMarkdown(bytes);

A path can still be supplied as a format hint when @mdgate/rtf is used through @mdgate/converters or a reader composed with @mdgate/core, but the path is never used to read a file from disk.

Compose it with other file readers

@mdgate/rtf implements the converter interface from @mdgate/core.

import { create } from '@mdgate/core';
import { rtf } from '@mdgate/rtf';
import { docx } from '@mdgate/docx';

const read = create([
  rtf(),
  docx(),
]);

The application still uses one reading interface while each format remains independently installable.

Need more than RTF?

If your application needs to read many different file types, use the complete converter set:

npm install @mdgate/converters

import { toMarkdown } from '@mdgate/converters';

const markdown = await toMarkdown(bytes, {
  path: filename,
});

@mdgate/rtf is one of the single-format packages in the open-source mdgate/converters project.

For AI agents, the same converter architecture can be used to extend read_file from text files to real-world document formats.

License

MIT