HTML DOM & RSS vs Accept: text/markdown & OKF
Rendered HTML5 DOM, CSS/JS bundles & RSS/Atom feeds is the classic web primitive built for human browsers and pre-integrated services (Render visual layout, typography, and interactive widgets for a human in a browser viewport), whereas Markdown for Agents and its step 4 (Read) peers provide the machine-native agentic surface (Accept: text/markdown) so autonomous agents can answer: "Can an agent ask a site for a Markdown version of an HTML page, instead of parsing the HTML itself?" Across the agenticweb.wiki conformance panel, Markdown for Agents is shipping (136 of 136 probed origins (last checked 2026-10-23)).
How HTML DOM & RSS compares to Accept: text/markdown & OKF at step 4 (Read): why the classic primitive breaks for software agents, and how sites deploy both together.
| Dimension | Rendered HTML5 DOM, CSS/JS bundles & RSS/Atom feeds Classic web | Markdown for Agents Agentic web |
|---|---|---|
| Primary consumer | Human driving a web browser or pre-integrated developer code | Autonomous software agent, coding harness, or AI search/RAG crawler |
| Lifecycle step | Classic web baseline for step 4 (Read) | 4 Read: Give me the content |
| Core question answered | Render visual layout, typography, and interactive widgets for a human in a browser viewport. | Can an agent ask a site for a Markdown version of an HTML page, instead of parsing the HTML itself? |
| Wire format & entrypoint | Accept: text/html (with navigation chrome, scripts, and inline SVG) | HTTP header at Accept: text/markdown |
| Failure mode for agents | Raw HTML pages typically spend 75–95% of their byte payload on layout boilerplate, navigation bars, scripts, and hydration state—wasting context tokens and burying tables or specifications deep in the byte stream. | Solved natively: Markdown for Agents is a Cloudflare network feature: when an agent requests a page with Accept: text/markdown, an enabled zone fetches the HTML from the origin, converts it to Markdown at the edge and returns it as text/markdown with Vary: Accept, so caches keep the HTML and Markdown variants apart. |
| Governance & status | Established IETF / W3C / industry baseline | Cloudflare; a feature of its network documented in Cloudflare's developer docs, not submitted to a standards body; it rides on ordinary HTTP content negotiation — Vendor feature, available to Pro, Business and Enterprise plans and SSL for SaaS customers at no cost; not a standards-track document |
| Live adoption (/check/ census) | Universal across the classic web | shipping — 136 of 136 probed origins (last checked 2026-10-23) |
| Primary specification | https://html.spec.whatwg.org/ | https://developers.cloudflare.com/fundamentals/reference/markdown-for-agents/ |
Why HTML DOM & RSS breaks when the caller is an agent
Raw HTML pages typically spend 75–95% of their byte payload on layout boilerplate, navigation bars, scripts, and hydration state—wasting context tokens and burying tables or specifications deep in the byte stream.
When to use which (and how they work together)
- Keep Rendered HTML5 DOM, CSS/JS bundles & RSS/Atom feeds: A human visitor opens your URL in Chrome, Safari, or Firefox.
- Add Markdown for Agents: An AI search crawler (Exa, Parallel.ai, OAI-SearchBot, ClaudeBot) or coding agent requests the same URL with Accept: text/markdown (or fetches <page>/index.md or /okf/index.md) to ingest clean, token-dense Markdown with Vary: Accept and x-markdown-tokens.
All agentic web standards replacing or extending HTML DOM & RSS
- Markdown for Agents — shipping · 136/136 probed origins: Can an agent ask a site for a Markdown version of an HTML page, instead of parsing the HTML itself?
- Open Knowledge Format (OKF) — shipping · 1/572 probed origins: How can an organization package durable, structured knowledge (schemas, metrics, playbooks, runbooks) so both humans and agents can read, produce, and trust it without a bespoke SDK?
- AGENTS.md — multi-implementation · 32/574 probed origins: What build, test, and convention instructions should a coding agent follow in this repository?
- Agent Readability Spec (ARS) — shipping: How readable is this site's content for AI agents, and what specifically should be fixed?
- Apache Ossie (formerly Open Semantic Interchange) (OSI) — prototype: How can an AI agent or BI tool read a metric or data-model definition and get the same meaning another tool intended?
- The 'knowledge-linkset' Well-Known URI for Publishing Knowledge Artefacts — none found: How can a web origin publish, at one well-known location, a checkable list of the knowledge artefacts it offers (graph serializations, JSON-LD context, text file, chunk export, change ledger) so that agents can find them and verify digests?