An agent-readable blog
Read public posts as Markdown and find blog feeds and discovery files.
Your published blog serves both HTML and Markdown. Drafts and private previews are not part of the public discovery files.
| URL or response | What an agent gets |
|---|---|
/llms.txt |
The blog name, summary, and links to published posts as Markdown. The list is bounded to 200 posts. |
/<slug>.md |
The published version’s Markdown with title, description, date, tags, and canonical URL in frontmatter. |
/<slug> with Accept: text/markdown |
The same Markdown representation. ?format=md also requests Markdown. |
/robots.txt |
Content-Signal: search=yes, ai-input=yes, ai-train=no for tenant blogs. |
/sitemap.xml |
Published post URLs when the blog is indexable. |
/feed.xml |
An RSS feed of published posts, bounded to 50 full-body items. |
HTML and Markdown article responses use Vary: Accept and an alternate Link header for the .md URL. A missing Markdown page returns a Markdown 404 instead of an HTML page. The blog home page also sends Link headers for llms.txt, the sitemap, and RSS. Blog pages include JSON-LD for the site and article.
curl https://your-blog.example/llms.txt
curl https://your-blog.example/example-post.md
curl -H 'Accept: text/markdown' https://your-blog.example/example-post
curl https://your-blog.example/robots.txt
curl https://your-blog.example/feed.xml
On a free default-domain blog, public pages remain readable but carry noindex; the sitemap is unavailable. A paid or self-hosted blog can be indexed.
Product discovery
The vibecms product and app hosts also publish discovery files for connecting an agent: an MCP server card at /.well-known/mcp/server-card.json, an API catalog at /.well-known/api-catalog, agent skill files under /.well-known/agent-skills/, and auth.md. Use those files on vibecms.dev or the app host to discover the API. Your blog’s files describe its published content.