Knowledge Layer vs. Cloudflare Markdown for Agents

The same standard for AI agents — on any site you run.

Cloudflare’s Markdown for Agents made a great idea concrete: serve every page as clean Markdown so AI agents read it without parsing HTML. Seogence speaks that exact standard — the same content negotiation, the same headers, the same frontmatter and embedded JSON-LD — and brings it to any host, with the schema.org structured data generated for you.

Get started See your page as Markdown

Why it’s an easy choice

One standard. Every host. Schema included.

Same open standard
Accept: text/markdown negotiation, the content-signal and token-count headers, and the frontmatter + JSON-LD layout. Agents that read a Cloudflare site read yours.
Any host, one snippet
Not just sites behind Cloudflare. Drop one embed on any stack — WordPress, Shopify, Wix, Next.js, or plain HTML — and the Knowledge Layer is live.
Schema generated, not just preserved
Cloudflare passes through the JSON-LD already on the page. Seogence writes the schema.org structured data for you, then embeds it in the Markdown.

Side by side

How they compare.

Capability	CloudflareMarkdown for Agents	SeogenceKnowledge Layer
Where it works Seogence extends	Sites proxied through Cloudflare (Pro, Business, Enterprise, and SSL for SaaS).	Any host — one embed snippet on WordPress, Shopify, Wix, Next.js, React, or plain HTML.
Request convention Same standard	Accept: text/markdown on the page URL.	Accept: text/markdown on the page-stable resolver, plus content-addressed .md and an llms.txt index.
Response headers Same standard	content-type: text/markdown, x-markdown-tokens, x-original-tokens, content-signal.	The same headers — including the same content-signal default (ai-train=yes, search=yes, ai-input=yes).
Output format Same standard	YAML frontmatter → clean body → JSON-LD in one fenced code block.	The same order, with richer frontmatter (url, author, site, keywords, dates).
Open Knowledge Format (OKF) Seogence extends	Plain frontmatter only — no OKF typing.	Typed frontmatter conformant with Google’s Open Knowledge Format (v0.1 draft), plus a typed organization.md node in the bundle.
Structured data Seogence extends	Preserves any schema.org JSON-LD already present in the page.	Generates the schema.org JSON-LD for each page, then embeds it in the Markdown.
How it is produced Seogence extends	Converted from live HTML on the fly at the edge.	Pre-generated and cached at content-addressed URLs; refreshed by content hash when the page changes.
Discovery Seogence extends	An llms.txt documentation index.	A per-domain llms.txt plus an auto-injected <link rel="alternate"> so non-JS crawlers find it too.
Source limits Seogence extends	HTML only, with a 2 MB origin-response cap.	Served from the captured snapshot — not bound by a per-request origin size cap.
Index hygiene Seogence extends	Not specified.	X-Robots-Tag: noindex on page copies, an indexable llms.txt — no duplicate-content risk.
Cost & control Seogence extends	Included at no extra cost on Cloudflare paid plans.	Opt-in per domain, toggled from your Seogence dashboard.

Parity rows reflect Cloudflare’s published “Markdown for agents” reference and Seogence’s shipped Knowledge Layer. Where the two match, agents need no special handling for either.

The same request

If you’ve called one, you’ve called the other.

The content-negotiation contract is identical. Point the request Cloudflare documents at your Seogence Knowledge Layer and you get the same content type, the same content-signal, and the same token-count headers so an agent can measure the savings before it spends a token.

Fetch a page as Markdown

# The exact content-negotiation request Cloudflare documents —
# pointed at your Seogence Knowledge Layer instead of an edge proxy.
curl "https://seogence.com/KnowledgeLayer/yoursite.com/by-url?u=https://yoursite.com/page" \
  -H "Accept: text/markdown"

# -> 200  content-type: text/markdown; charset=utf-8
#         content-signal: ai-train=yes, search=yes, ai-input=yes
#         x-markdown-tokens: 342   x-original-tokens: 5188

Same idea, two architectures

An edge proxy vs. a universal embed.

The difference isn’t the output — it’s how it reaches the agent. Cloudflare can negotiate Markdown on your own URL because it sits in front of your origin. Seogence isn’t in your request path, so it serves the Markdown from the Knowledge Layer and points agents to it — which is exactly why it works on hosts that never touch Cloudflare.

Cloudflare — at the edge

Your site is proxied through Cloudflare. A request with Accept: text/markdown to the normal page URL is converted from live HTML and returned in place. Requires the site to run on Cloudflare’s paid plans.

Seogence — via the embed

One snippet, any host. Seogence generates the Markdown (with the schema.org it writes), serves it from the Knowledge Layer, and advertises it with a per-domain llms.txt and an injected <link rel="alternate"> — discoverable even without JavaScript.

Questions

Common questions.

Is Seogence compatible with Cloudflare’s Markdown for Agents?

Yes. Seogence serves the same Accept: text/markdown content negotiation, the same content-type, content-signal and token-count headers, and the same frontmatter → body → JSON-LD layout. An agent built to read a Cloudflare site reads a Seogence-powered site the same way.

Do I need to be on Cloudflare to use Seogence?

No. Cloudflare’s feature only covers sites proxied through its paid plans. Seogence’s Knowledge Layer works on any host — you add one embed snippet, on WordPress, Shopify, Wix, Next.js, React, or plain HTML.

What does Seogence add beyond converting HTML to Markdown?

Seogence generates the schema.org structured data for each page and embeds it in the Markdown, adds a per-domain llms.txt index and an auto-injected discovery link for non-JS crawlers, serves content-addressed URLs, and keeps every copy fresh by content hash.

Can I run both together?

Yes. They follow the same convention, so running Cloudflare’s edge conversion and Seogence’s Knowledge Layer side by side is harmless. Seogence adds the schema generation and works across any host you run.

What does the Open Knowledge Format (OKF) row mean?

Seogence’s markdown carries typed frontmatter conforming to Google Cloud’s Open Knowledge Format (v0.1 draft), and the bundle includes a typed organization.md node; Cloudflare’s output uses plain frontmatter without OKF typing. To be clear about what OKF buys you: it is a draft, no public crawler consumes OKF bundles today, and Google says it has no effect on Search ranking — it is forward-compatible typing of the same markdown, with nothing replaced.

Want the full reference? See Markdown for AI agents for the URL patterns, headers, and output format in detail.

Try it

See your page as Markdown — the way an AI reads it.

Enter any live page. Seogence analyzes it and shows the clean Markdown copy — with the structured data it would add — so you can see exactly what agents would consume, on any host.