---
title: "Markdown (page.md)"
description: "Your page, in the format LLMs were trained on."
canonical: "https://geordy.ai/formats/markdown"
---
# Markdown

**Your page, in the format LLMs were trained on.**

Clean Markdown stripped of nav, ads, scripts, and CSS. Headings preserved, link targets normalized, content placed in a single retrieval-tight document. The format AI reads first when given a choice.

Markdown is a lightweight markup language with plain-text formatting syntax designed to be converted to HTML and many other formats. Created in 2004 by John Gruber with Aaron Swartz, it was designed to enable people to write using an easy-to-read and easy-to-write plain text format that could be converted to structurally valid HTML.

## At a glance

- File: `page.md`
- First released: 2004
- Created by: John Gruber & Aaron Swartz
- Specification: https://spec.commonmark.org/current/
- Read by: GPTBot, ClaudeBot, PerplexityBot, General LLM systems

## Why it matters for AI

Markdown is the lowest-friction text format LLMs are trained on at massive scale - model weights understand `#`, `>`, `**` and bullet lists natively, so well-structured Markdown is parsed more reliably than HTML when ingested into context windows or RAG pipelines. Almost every AI coding tool, doc site, and chat UI consumes it as first-class input.

## Example

```
# Geordy

## What is GEO?

GEO is the practice of optimizing content for AI-powered search engines.

## Key Features

- **Multi-format generation**: 16 AI-optimized formats
- **AI bot tracking**: Monitor GPTBot, ClaudeBot, PerplexityBot
```

## Benefits

- Simple, intuitive syntax that's easy to learn
- Widely supported across major platforms (GitHub, Reddit, Stack Overflow)
- Plain text format compatible with version control
- No special software required - works in any text editor

## Limitations

- Limited formatting compared to HTML
- Inconsistent implementations across platforms
- No standard for complex tables or diagrams

## Best practices

- Use ATX-style headings (`#`, `##`) and keep a single H1 per document; LLMs use heading hierarchy to chunk and route content.
- Prefer fenced code blocks with a language tag (```python) - improves both human syntax highlighting and LLM code understanding.
- Write descriptive link text - `[CommonMark spec](https://spec.commonmark.org)` not `[click here](...)`; link text becomes context for the model.
- Stick to CommonMark or GitHub Flavored Markdown; avoid renderer-specific extensions if the file will be ingested by external tools.

## Pitfalls

- Mixing tabs and spaces in nested lists - most parsers and LLM tokenizers lose the structure.
- Skipping a blank line before a list, code block, or heading - the original Markdown spec requires it and many parsers will silently merge the elements.
- Embedding raw HTML for layout (divs, spans) - defeats portability and often gets stripped in RAG pre-processing.

## Use cases

- **Documentation and README files**
- **Content writing and blog posts**
- **Note-taking and knowledge management**

## Adopters

- GitHub (https://github.com)
- Stack Overflow (https://stackoverflow.com)
- Reddit (https://www.reddit.com)
- Discord (https://discord.com)
- Obsidian / Notion (https://obsidian.md)
---

Source: https://geordy.ai/formats/markdown
This is a machine-readable markdown version of that page, generated by Geordy.