In development · First edition 2026

An independent, open-data read on how the web is adapting to AI agents.

Most numbers about the agentic web come from vendors measuring their own tools. We're building the independent one, sourced from public datasets, and a handful of founding sponsors will help bring the first edition to life.

Published by Aglarond · llms.txt · MCP · structured data · AI-crawler rules
Why this exists

The agentic web is being built. Nobody neutral is measuring it.

llms.txt, MCP, agent cards, structured data, AI-crawler permissions: adoption is moving fast, but every report is published by a company about its own product. There's no independent baseline. We're building one from public data.

A sample of the data

An example: llms.txt adoption.

The kind of signal the report tracks, sourced from public datasets. Here's how valid llms.txt has spread across the top 10,000 sites.

~5×Growth in valid llms.txt across the top 10,000 sites since mid-2025, to just under 5%.

One of many signals the report will track across the agentic web.

Chart: llms.txt adoption across the top 10,000 sites, rising from about 0.9% to about 4.7% between July 2025 and May 2026.
Illustrative example, sourced from public datasets. Top 10,000 sites by Chrome UX Report popularity.
The working scope

What we're planning to measure.

An early look at the themes we're building toward. Each signal will be graded by how we can measure it: web-scale, sampled, or explained where it can't yet be counted.

Access & permissions

Who the web lets in: AI-crawler rules in robots.txt, bot-defense walls, the open-vs-blocked tension.

Agent content interfaces

llms.txt, machine-readable feeds, and whether a page is even readable without running JavaScript.

Structured data & discovery

Schema.org coverage, sitemaps, link headers: the typed facts agents extract.

The agent-protocol frontier

NLWeb, WebMCP, agent cards, MCP: the newest layer, mostly still aspirational.

Agent-readiness in practice

What "agent-ready" actually looks like, and who's doing it.

Are the agents showing up?

The demand side: AI-crawler traffic and how it's trending.

Working scope: themes and framing are still being finalized.

Why sponsor now

Back the category-definer before it's crowded.

Own the association

Sponsors of the first credible, independent report in a young category get a positioning advantage that's far cheaper now than later.

Open by default

Sourced from public datasets and sources. No tool to sell, no vendor agenda.

Founding terms

A handful of slots, founding pricing, and a co-branded data feature with full credit. First in, most visible.

Why you can trust it

Independent by construction.

No tool to sell

We're not a vendor. We measure the whole web, not our own slice of it.

Public sources

Built on public datasets and sources.

Independent

Published by Aglarond, an independent studio.

Who's behind it
Casey Burridge

Casey Burridge

Strategic Growth & Operations Manager, GravityKit · Founder, Aglarond

Six-plus years building growth for WordPress and SaaS companies, with a Master's in Information Management. Casey works at the overlap of marketing, data, and AI-first ops.

The Agentic Web Report is original-data research, not commentary. It's published by Aglarond, an independent studio built on one idea: the most defensible thing you can do in this space is measure it honestly, in the open, and let the data speak.

Get involved

Be part of the first edition.

Founding-sponsor slots are open for the 2026 edition. Grab the sponsor pack, or get in touch.