AI & Language Infrastructure

MarkItDownJS

Universal Document-to-Markdown Conversion Engine

AST-first document ingestion and conversion platform for TypeScript. Converts PDFs, Office docs, images, audio, and archives into structured AI-ready Markdown with RAG chunking.

CodeAll Projects
TypeScriptNode.jsTurborepopnpmVite

Project Facts

NameMarkItDownJS
CategoryAI & Language Infrastructure
Primary LanguageTypeScript
TechnologiesTypeScript, Node.js, Turborepo, pnpm, Vite
Sourcehttps://github.com/instax-dutta/MarkItDownJS

More in AI & Language Infrastructure

Ansora

Self-Hostable Serverless Blogging Platform

Visitor Analytics

Privacy-Preserving Analytics SDK

What is MarkItDownJS?

MarkItDownJS is an open-source project by Sai Dutta Abhishek Dash, a founder and engineer based in Odisha, India. It belongs to the AI & Language Infrastructure category, and is a Universal Document-to-Markdown Conversion Engine. AST-first document ingestion and conversion platform for TypeScript. Converts PDFs, Office docs, images, audio, and archives into structured AI-ready Markdown with RAG chunking. The primary implementation language is TypeScript. It uses TypeScript, Node.js, Turborepo, pnpm, Vite. Handles PDFs, DOCX, PPTX, XLSX, HTML, CSV, JSON, XML, images, audio, EPUB, and archives. Built as a Turborepo monorepo with RAG-ready chunking applied to the output. Other systems in the same track are ornith-flight, AgentLoop, PhishScout, and Hinglish-Bench. The full source code is public at https://github.com/instax-dutta/MarkItDownJS. Sai Dutta Abhishek Dash has shipped 22 open-source production systems, maintains 85+ public repositories on GitHub, and has published 46+ models and 22+ datasets on Hugging Face.