About — LLM-LD

Why LLM-LD Exists

Search is changing. AI systems — ChatGPT, Gemini, Perplexity, Claude — are becoming the way people find information, choose businesses, and make decisions. But these systems can't reliably understand most websites.

Traditional SEO was built for Google's index. It optimizes for keywords and links. AI systems need something different: structured, machine-readable data that tells them exactly what a business does, who it serves, where it operates, and how to take action.

LLM-LD bridges that gap. It's a standard that gives any website a machine-readable AI layer — structured data that AI systems can consume, understand, and use to recommend your business in the right context.


Principles

📖

Open Standard

The spec is published under CC BY 4.0. Anyone can read it, implement it, extend it, or build tools for it. No permission needed.

🆓

Free Forever

No licensing fees to implement. No subscription to join the network. No premium tiers. The standard is free. The network is free.

🛡️

Defense in Depth

Three independent layers that each work on their own. Even if no AI platform ever formally adopts the standard, your site is still more AI-readable than any competitor's.

🔧

Built from Practice

Every part of LLM-LD was developed through real implementations across real client sites. The spec documents what works — it doesn't theorize.


Timeline

2024

Problem Identified

CAPXEL begins working with clients on AI search visibility. Discovers that no standard exists for structuring website data for LLMs. Begins developing internal frameworks.

Late 2024

First Implementations

Framework tested across first wave of client sites. The three-layer architecture — Schema.org injections, knowledge graph files, llm-index.json — proves effective in real AI search results.

Mid 2025

First Discovery Network Launches

Early discovery network goes live. First batch of sites joins the master sitemap. The "find one, discover all" model begins working in practice.

Feb 2026

LLM-LD 1.0 Published + LLM Disco goes live

The standard is formalized and published at llmld.org. Spec, guides, and examples made publicly available. 100+ sites in the network. Open for adoption.


Who Builds This

LLM-LD and the LLM Disco Network are created and maintained by CAPXEL, an AI data company specializing in AI audience targeting and AI search optimization (ASO).

CAPXEL develops the standard, operates the network, and implements LLM-LD for clients. The standard is open because a standard only works if anyone can use it. The network is free because a network only works if people join it.

There is no foundation, consortium, or committee. CAPXEL builds it, ships it, and iterates based on real-world implementation experience across the network. If adoption grows to warrant governance changes, that structure will be built — but only when it's needed, not before.


Get in Touch

Questions about the standard, the network, or implementation? We'd love to hear from you.