**Bloxx Pages**

  Running on 27+ live domains

# Publish a page from your AI in one line

The page is live in seconds, with title, description, canonical and JSON-LD already correct. Then it is served as .md, .txt, .json and .schema at the same URL, for anything that reads pages by machine.

  [How it works](#how)[Read the language spec](/bloxx-markdown)

  This page, in four machine formats
  [/home](/home)[.md](/home.md)[.txt](/home.txt)[.json](/home.json)[.schema](/home.schema)

Click any of them. Same content, same URL, served the way a crawler or an agent asks for it. Every page published through Bloxx gets this, cached at the edge.

27+live domains served by the Bloxx renderer today4machine formats at every URL, no configuration10tools in the connector, from publish to AI-crawler share

## The gap this fills

Netlify Drop, tiiny.host and Claude artifact sharing all solve hosting. None of them touch the page on the way through. You get a link, and the page ships with whatever meta, heading structure and structured data the generator happened to emit, which for AI-generated HTML is usually none.

Bloxx treats publishing as the moment to fix that, because it is the only moment where the fix is free. The page is already in a pipeline, so scoring it, repairing the head, and exposing it in machine formats costs the user nothing and no extra step.

Being precise about status, because it matters more than a tidy story: the publish loop, the format views, the audit engine, the analytics tools and Bloxx Markdown are live and in daily use across the fleet. Audit-on-publish, custom domains and lead capture are next, not done.

## What is live today

1. 1 ### Publish one connector call, or one paste of a full HTML document. The page is a live URL on bloxx.page in seconds, and republishing the same slug updates that page instead of making a duplicate.
2. 2 ### Score the audit engine reads any live URL and returns overall, schema, LLM-readability and semantics scores, with the specific findings behind each one.
3. 3 ### Measure every request to every Bloxx domain is classified at the edge, so AI crawlers show up separately from search crawlers, SEO tools and humans.
4. 4 ### Serve to machines .md, .txt, .json and .schema are generated from the same page and cached at the edge, so an agent takes the format it wants without scraping HTML.

## One connector for the whole site

The part that is genuinely hard to assemble yourself is not the hosting. It is having publishing, optimisation scoring, Search Console, GA4, AI referral traffic and crawler data available in one place, to one agent, in one session. That is the connector.

### publish_page

takes a full HTML document and returns a live URL. Idempotent on the slug, purges the edge cache on write.

### audit_site, compare_sites

schema, LLM-readability and semantics scores for any URL, with findings, quick wins and an action plan.

### search_console_performance

clicks, impressions, position and CTR from Search Console, by query, page, country or device.

### striking_distance

the queries sitting at positions 5 to 20, ranked by how much click opportunity is actually available.

### analytics_overview, traffic_sources

GA4 sessions, users and conversions, with an AI-referred segment split out.

### ai_referral_report, bot_traffic_report

what arrives from ChatGPT, Perplexity, Gemini and Copilot, and which AI crawlers are reading the site.

Point it at a domain and ask what is wrong with it. The answer comes back from the audit engine, Search Console, GA4 and the edge logs together, which is the part no single free MCP server does.

Add the connector*json*

```
{
  "mcpServers": {
    "bloxx": {
      "url": "https://mcp.bloxx.page/mcp"
    }
  }
}
```

## The page language

Prose is ordinary Markdown. Structure comes from directives that expand to pre-built, already-optimised HTML, and the structured ones emit their own JSON-LD. A pricing block becomes a Product with an Offer. An FAQ block becomes a FAQPage. Nobody writes schema by hand, and nobody forgets it.

The page you are reading, in Bloxx Markdown*bloxx-markdown*

```
:::hero{title="Publish a page from your AI in one line" cta="How it works" href="#how"}

:::formats{url="/home" label="This page, in four machine formats"}

:::faq
- Q: Is the page live instantly? A: Yes, in seconds.
- Q: Can an agent read it? A: Every page is also .md, .txt, .json and .schema.
:::
```

That is the whole idea. An agent writes a few lines instead of four hundred lines of HTML, and what comes out is consistent, styled and schema-correct because the components carry their own structured data. The [language spec](/bloxx-markdown) has the grammar and the full catalogue.

## Formats and modifiers

One URL grammar, applied the same way to every page: formats as a suffix, modifiers as a query. Learn it once, use it everywhere.

  Try it on another page
  [/bloxx-markdown](/bloxx-markdown)[.md](/bloxx-markdown.md)[.txt](/bloxx-markdown.txt)[.json](/bloxx-markdown.json)[.schema](/bloxx-markdown.schema)

The format views are not a per-page feature you switch on. They are a property of the renderer, so every page on every Bloxx domain has them. The [formats reference](/formats) covers the full grammar.

What does born optimised actually mean?

The head is correct on arrival: title, meta description, canonical, robots, Open Graph and a JSON-LD graph assembled from the page's own components, rather than left to whatever the generator emitted.

Do I have to use Bloxx Markdown?

No. A full HTML document publishes as-is and still gets the format views, the edge caching and the audit engine. The language is for when you want an agent to write pages cheaply and consistently.

How is this different from MDX or Markdoc?

The grammar is deliberately similar and that is the point, because models already write it. The difference is the compile target: an optimised page plus its JSON-LD plus machine formats plus hosting, rather than a React component tree you still have to build and deploy.

Why does the AI-crawler data matter?

AI referral traffic is still a small share of all traffic, but it converts better than most channels, and crawler visits are the leading indicator of it. Reading them at the edge is the earliest signal available that an answer engine has started paying attention.

Can I use my own domain?

Not yet. Pages publish under bloxx.page today; custom domains are the next piece of table stakes.

## Early access

The connector is live and the publish loop is in daily use across the fleet. If you build or manage sites and want to try it, get in touch.

  [Email hello@itsbloxx.com](mailto:hello@itsbloxx.com)[Run a free audit](https://itsbloxx.com/r/audit)