Introduction to Omni-Core

Last Updated March 31, 2026

Welcome to the official documentation for Omni-MDX, the high-performance, cross-language MDX engine powered by Rust.

The Vision: Solving Parser Fragmentation

Historically, processing rich text and complex components (MDX) meant relying on JavaScript engines like mdx-js. While powerful, these engines create a significant bottleneck when you need to parse massive amounts of data, or when you want to use the exact same logic outside of a Node.js ecosystem (like in a Python data pipeline or a native desktop app).

Instead of having one parser in JavaScript for your website, another in Python for your datasets, and yet another for mobile, Omni-Core introduces a radical approach: A Single Core in Rust.

By centralizing the parsing logic in a hyper-optimized Rust binary, we guarantee identical behavior, memory safety, and native execution speed across all your platforms.

Key Principles

Omni-MDX was built from the ground up with four core pillars in mind:

1. Blazing Fast Native Speed

Thanks to Rust’s zero-cost abstractions and safe memory management (using Arc reference counting instead of heavy garbage collection), Omni-MDX parses complex documents up to 30x faster than traditional JS parsers. Generating an Abstract Syntax Tree (AST) takes fractions of a millisecond.

2. Math-Ready by Default

Stop hacking standard Markdown parsers to support equations. Omni-MDX has native lexing for complex mathematical formulas. Whether it’s inline math ($E=mc^2$) or block equations, it parses them perfectly, making it the ideal engine for research papers and academic datasets.

3. Component Agnostic

Omni-MDX doesn’t care how you render your UI. Its job is to identify JSX/MDX components seamlessly and output a pristine AST. It delegates the actual rendering to the host platform. A <Chart /> tag can become a React component on the web, or a PyQt widget in a native desktop application.

4. Universal Multi-Target FFI

Write once, run everywhere. The single source code compiles down to:

• WASM for the Web (Next.js, React).
• PyO3 for Python (Data processing, AI pipelines, PyQt).
• Coming soon: Dart/FFI for mobile.

Architecture Overview

Under the hood, the Omni-Core monorepo is divided into three main layers:

1. core-parser/: The Rust heart. Contains the Lexer, Parser, Binary Encoder, and AST definitions.
2. packages/mdx-next/: The Next.js integration. Decodes the binary AST into JavaScript objects without serialization overhead.
3. packages/mdx-python/: The Python engine. Exposes the parser to Python environments for backend data ingestion or Qt rendering.

Ready to see it in action? Head over to the Installation Guide to get started.