--- title: Markdown Became the Control Plane for AI | daily.dev description: Markdown, born as a Perl script in 2004, has quietly become the control plane for AI systems. Agent instruction files, skill definitions, and model training... canonical: https://daily.dev/posts/markdown-became-the-control-plane-for-ai-afcgdibtk twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Markdown Became the Control Plane for AI | daily.dev og:description: Markdown, born as a Perl script in 2004, has quietly become the control plane for AI systems. Agent instruction files, skill definitions, and model training... og:url: https://daily.dev/posts/markdown-became-the-control-plane-for-ai-afcgdibtk og:image: https://api.daily.dev/og/posts/AFCGdibtK.png og:image:alt: Markdown Became the Control Plane for AI og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Markdown Became the Control Plane for AI **[API Evangelist](https://daily.dev/sources/apievangelist)** · 6 min read · 23 upvotes · 2 comments ## Summary Markdown, born as a Perl script in 2004, has quietly become the control plane for AI systems. Agent instruction files, skill definitions, and model training corpora all rely on it. The post explores why markdown won (graceful degradation of meaning, not just simplicity), the fragmented landscape of named markdown conventions (AGENTS.md, CLAUDE.md, SKILL.md, llms.txt, etc.), and real adoption data from 26,099 API providers in the APIs.io catalog. Key findings: 24% publish llms.txt, 9.6% declare agent skills, but AGENTS.md appears at only 0.4% of commercial providers. The core recommendation is to serve markdown as a representation via content negotiation, not just use it as a source format. The post promotes a $50 paper, 'The Fundamentals of Markdown,' with supporting data and analysis. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What percentage of API providers actually publish an llms.txt file? About 24% (6,260 out of 26,099 providers cataloged) publish a fetchable llms.txt file, which is faster uptake than OpenAPI achieved in its first two years. However, only 31 of those providers ship the full-corpus companion file, meaning most publish just an index and stop there. _daily.dev surfaces adoption data like this for teams deciding whether llms.txt is worth prioritizing._ ### How common is AGENTS.md adoption among commercial API providers versus open source projects? AGENTS.md appears in only about 0.4% of commercial API providers' published surfaces, despite being used in over 60,000 open-source projects. The gap exists because providers tend to treat agent instructions as an internal codebase concern rather than something to publish externally to API consumers. _teams weighing whether to expose agent instruction files publicly can track conventions like AGENTS.md via daily.dev._ ### Should I use .md or .mdx for content that AI agents will read? Plain .md is overwhelmingly dominant for agent-readable content, outnumbering .mdx by roughly 257,980 to 269 files across a large catalog of API-related markdown. For anything an agent might read, plain markdown without JSX extensions is the safer, more widely supported choice. _developers picking a documentation format for AI tooling can follow format trends like this on daily.dev._ ## Community discussion Top comments from developers on daily.dev. **@kartiknvj** · 0 upvotes > "Graceful degradation of meaning" is a sharper explanation than simplicity, and it matches what I see: agents recover from a half-broken Markdown instruction file far better than from malformed JSON. The 0.4% [AGENTS.md](http://AGENTS.md) number is the real story though, the conventions are proliferating faster than anyone is adopting a single one. Do you think content negotiation actually converges them, or do we just end up with more competing filenames? **@agustinbarrientos** · 0 upvotes > The CDN cache key has to vary on Accept or a Markdown response can reach a browser. ## Similar posts on daily.dev - [Andrej Karpathy, Google and Garry Tan agree Markdown is the answer, but they’re not solving the same problem](https://daily.dev/posts/andrej-karpathy-google-and-garry-tan-agree-markdown-is-the-answer-but-they-re-not-solving-the-same-lxq4tg8i0) · The New Stack · 0 upvotes · 0 comments - [Serving Markdown for AI Agents](https://daily.dev/posts/serving-markdown-for-ai-agents-j2txtbg8z) · RUBYLAND · 2 upvotes · 1 comments - [Agents grew up, so did our docs](https://daily.dev/posts/agents-grew-up-so-did-our-docs-rucdfhfyl) · Neon · 0 upvotes · 0 comments - [Why markdown is becoming the default language between search data and AI models](https://daily.dev/posts/why-markdown-is-becoming-the-default-language-between-search-data-and-ai-models-qtysxpc0p) · The Next Web · 0 upvotes · 0 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#markdown](https://daily.dev/tags/markdown) [View this post on daily.dev](https://daily.dev/posts/markdown-became-the-control-plane-for-ai-afcgdibtk) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Markdown Became the Control Plane for AI","url":"https://daily.dev/posts/markdown-became-the-control-plane-for-ai-afcgdibtk","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/markdown-became-the-control-plane-for-ai-afcgdibtk"},"datePublished":"2026-08-10T12:32:07.817Z","dateModified":"2026-09-13T19:43:53.274Z","description":"Markdown, born as a Perl script in 2004, has quietly become the control plane for AI systems. Agent instruction files, skill definitions, and model training...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/3691aa9599120582c07b1a2c230d09bb?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/3691aa9599120582c07b1a2c230d09bb?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"API Evangelist","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"API Evangelist","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/a36244ae67bc4f41a605f780267ecb5f","url":"https://daily.dev/sources/apievangelist"},"commentCount":2,"discussionUrl":"https://daily.dev/posts/markdown-became-the-control-plane-for-ai-afcgdibtk","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":23},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":2}],"keywords":"ai-agents,markdown","timeRequired":"PT6M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"API Evangelist","item":"https://daily.dev/sources/apievangelist"},{"@type":"ListItem","position":3,"name":"Markdown Became the Control Plane for AI"}]} {"@context":"https://schema.org","@type":"WebPage","@id":"https://daily.dev/posts/markdown-became-the-control-plane-for-ai-afcgdibtk","comment":[{"@type":"Comment","text":"“Graceful degradation of meaning” is a sharper explanation than simplicity, and it matches what I see: agents recover from a half-broken Markdown instruction file far better than from malformed JSON. The 0.4% AGENTS.md number is the real story though, the conventions are proliferating faster than anyone is adopting a single one. Do you think content negotiation actually converges them, or do we just end up with more competing filenames?","datePublished":"2026-08-12T14:12:16.685Z","url":"https://daily.dev/posts/AFCGdibtK#c-7qd6myaEe","author":{"@type":"Person","name":"kartik-nvjk","url":"https://daily.dev/kartiknvj","image":"https://media.daily.dev/image/upload/s--3gGgsVCw--/f_auto/v1781456774/avatars/avatar_TvTVeiMdkRCqWUDullFmy?_a=BAMAMiWQ0"}},{"@type":"Comment","text":"The CDN cache key has to vary on Accept or a Markdown response can reach a browser.","datePublished":"2026-08-15T22:20:26.095Z","url":"https://daily.dev/posts/AFCGdibtK#c-5Za68A6eg","author":{"@type":"Person","name":"Agustin Barrientos","url":"https://daily.dev/agustinbarrientos","image":"https://media.daily.dev/image/upload/s--5ayxQnqn--/f_auto/v1788281802/avatars/avatar_wQYYVe5Tbj0NJ7C7qPoa8?_a=BAMAMicg0"}}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/markdown-became-the-control-plane-for-ai-afcgdibtk#faq","mainEntity":[{"@type":"Question","name":"What percentage of API providers actually publish an llms.txt file?","acceptedAnswer":{"@type":"Answer","text":"About 24% (6,260 out of 26,099 providers cataloged) publish a fetchable llms.txt file, which is faster uptake than OpenAPI achieved in its first two years. However, only 31 of those providers ship the full-corpus companion file, meaning most publish just an index and stop there. daily.dev surfaces adoption data like this for teams deciding whether llms.txt is worth prioritizing."}},{"@type":"Question","name":"How common is AGENTS.md adoption among commercial API providers versus open source projects?","acceptedAnswer":{"@type":"Answer","text":"AGENTS.md appears in only about 0.4% of commercial API providers' published surfaces, despite being used in over 60,000 open-source projects. The gap exists because providers tend to treat agent instructions as an internal codebase concern rather than something to publish externally to API consumers. teams weighing whether to expose agent instruction files publicly can track conventions like AGENTS.md via daily.dev."}},{"@type":"Question","name":"Should I use .md or .mdx for content that AI agents will read?","acceptedAnswer":{"@type":"Answer","text":"Plain .md is overwhelmingly dominant for agent-readable content, outnumbering .mdx by roughly 257,980 to 269 files across a large catalog of API-related markdown. For anything an agent might read, plain markdown without JSX extensions is the safer, more widely supported choice. developers picking a documentation format for AI tooling can follow format trends like this on daily.dev."}}]} ```