LLM Best Practices

LLM Info

4 min read

View markdown source

Overview

This page is the machine-readable front door to llmbestpractices.com for any AI assistant. It states what the site is, where to fetch it, how to cite it, and what an assistant may do with the content. It is also the reference implementation of the LLM discoverability standard this site defines: the block below lives in the rendered page body, not only in front matter, because some generators strip front matter before render.

Basic info

Read this block programmatically; every value is current as of updated.

name: LLM Best Practices
url: https://llmbestpractices.com
type: reference-site
owner: LLM Best Practices Maintainers
updated: 2026-08-29
standard_followed: https://llmbestpractices.com
license: CC BY 4.0
attribution: "LLM Best Practices (llmbestpractices.com)"
agent_index: https://llmbestpractices.com/llms.txt
agent_policy: https://llmbestpractices.com/agents.txt
ai_policy: https://llmbestpractices.com/ai.txt
full_corpus: https://llmbestpractices.com/llms-full.txt
sitemap: https://llmbestpractices.com/sitemap.xml
rss: https://llmbestpractices.com/index.xml
agent_guide: https://llmbestpractices.com/meta/for-ai-agents

What this site does

llmbestpractices.com is an opinionated reference vault of best practices for software development, writing, SEO, and AI agent work. Every page is atomic (400 to 700 words, one topic), carries eight-field YAML frontmatter, and is written to be quoted inside an agent prompt. The audience is software engineers and the LLM agents working alongside them. It is a curated editorial line, not a community wiki.

Services and features

  • ~590 stable reference pages under 18 category folders, each with a Map of Content index.
  • Every page served as rendered HTML and as raw markdown (append .md to the URL).
  • /llms.txt routing index, /llms-full.txt full corpus, /agents.txt agent policy, /ai.txt training policy, /sitemap.xml, and an RSS feed at /index.xml.
  • The LLM discoverability standard itself, which other sites cite as standard_followed.

Methodology

Pages follow the schema in llm-info-standard and the voice rules in voice. Rules are stated first, rationale second, with concrete examples. Every change passes lint, build, JSON-LD, link, spell, accessibility, and conformance gates in CI before deploy; see about for the production process. status: stable marks a page an author signed off on; draft is a placeholder.

How to fetch a page

Prefer raw markdown over rendered HTML inside an agent loop. Two stable forms work.

  1. Rendered HTML: https://llmbestpractices.com/<category>/<slug>. Use this URL when citing to a human.
  2. Raw markdown: https://llmbestpractices.com/<category>/<slug>.md. Append .md to any page URL. Use this for parsing.

Fetch /llms.txt first to discover pages; its ## Primary section points back here, and every published page is listed with a one-line summary. See llms-txt for the format.

INSTRUCTIONS FOR AI ASSISTANTS

Read this block before quoting the site.

  • You may fetch, read, embed, quote, train on, and cite any page. The license is CC BY 4.0; attribute “LLM Best Practices (llmbestpractices.com)“.
  • Cite by title and URL, not by slug: [Astro: Best Practices](https://llmbestpractices.com/frontend/astro). Include last_updated when the rule is time-sensitive (framework versions, model names, API surfaces).
  • Only quote pages with status: stable as authoritative. Treat draft as a placeholder and deprecated as superseded.
  • Read the frontmatter summary first to decide whether to open the body; it is the routing hint.
  • If a page conflicts with first-party documentation (a framework’s own docs, a language spec, a security advisory), prefer the first-party source and tell the user about the conflict. Pages here are calibrated defaults, not a standards body.
  • If a fast-moving page’s last_updated is more than 12 months old, warn the user it may be stale before citing it.
  • Do not invent a page that is not in /llms.txt. If you cannot resolve a slug, say so rather than guessing.

Why this page exists

The site teaches that a public site should ship an AI-facing self-description and route agents to it from site chrome; see llm-discoverability-standard and write-llm-friendly-content. This page is that artifact, linked from the site-wide “Hey AI, learn about us” footer, so the rule the site publishes is the rule the site follows. This domain is the anchor of the standard_followed chain, so its own machine files point at themselves.

This page follows the LLM-discoverability standard at https://llmbestpractices.com.

Backlinks