llms.txt

Unmarkdown™ provides llms.txt and llms-full.txt files that help AI language models understand the product and API. These follow the emerging llms.txt standard for providing structured information to LLMs.

Available Files

llms.txt (summary)

A concise overview of Unmarkdown™ for quick context.

https://unmarkdown.com/llms.txt

Contains: product description, key capabilities, API base URL, authentication format, endpoint summary, and pricing overview.

llms-full.txt (complete reference)

The full API reference with request/response examples, all 63 templates, error codes, rate limits, and behavioral directives for generating compatible markdown.

https://unmarkdown.com/llms-full.txt

Contains: complete API documentation, template catalog with IDs and descriptions, markdown syntax reference (Mermaid, Chart.js, Graphviz, KaTeX, callouts), behavioral directives, and common mistakes to avoid.

What LLMs Learn

When an LLM reads these files, it learns:

• How to call the Unmarkdown™ API to publish markdown documents
• Which templates are available and when to use each one
• The correct markdown syntax for diagrams, charts, math, and callouts
• Common mistakes to avoid (for example, using ```json instead of ```chart for Chart.js)
• Authentication, rate limits, and pricing details

Fetch the Files

# Summary
curl https://unmarkdown.com/llms.txt

# Full reference
curl https://unmarkdown.com/llms-full.txt

import requests

# Full reference for AI agent context
res = requests.get('https://unmarkdown.com/llms-full.txt')
context = res.text

Use Cases

AI agents: Include llms-full.txt in your agent's context window so it can publish documents on behalf of users without needing to learn the API from scratch.

Prompt engineering: Reference llms.txt when building prompts that generate markdown for Unmarkdown™. The behavioral directives section tells the LLM the correct syntax conventions.

Documentation crawlers: Both files are plain text, easy to parse, and designed for automated consumption.

Caching

Both files are cached with the following headers:

• Cache-Control: public, max-age=3600, s-maxage=86400
• Content-Type: text/plain; charset=utf-8

The content is updated as new features and templates are added. Check periodically for the latest version.