The GEO and Agentic Web Glossary
The vocabulary of AI search, defined and then actually implemented. Every entry gives you a short definition, links to the primary spec, three working implementation examples, and the guidelines that stop it going wrong in production.
Agent & AI protocols
11 termsThe standards that let an AI agent do something on your site rather than just read it. Most of these landed between 2025 and 2026 and adoption is still low, which is exactly why implementing them early is a visibility advantage.
WebMCP
Web Model Context Protocol
Let AI agents call your site's own tools in the browser instead of guessing at your UI.
Read the entry →MCP
Model Context Protocol
The open standard that connects AI models to your APIs, data and tools server-side.
Read the entry →UCP
Universal Commerce Protocol
Expose catalog, cart and checkout to shopping agents through a signed, discoverable profile.
Read the entry →A2A
Agent2Agent Protocol
Let independent agents delegate tasks to each other across organisational boundaries.
Read the entry →Agent Card
/.well-known/agent-card.json
The machine-readable business card that tells agents what your service can do and how to call it.
Read the entry →AP2
Agent Payments Protocol
Signed mandates that let an agent pay on a user's behalf without holding their card.
Read the entry →Agent Skills
Agent Skills (SKILL.md)
Packaged, discoverable instructions that teach an agent how to work with your product.
Read the entry →NLWeb
Natural Language Web
Turn your existing structured data into a natural-language endpoint agents can query.
Read the entry →Schemamap
Schema mapping for agents
Publish how your internal data model maps to public vocabulary so agents can query it correctly.
Read the entry →Web Bot Auth
HTTP Message Signatures (RFC 9421)
Cryptographically prove which bot is knocking, so you can allow good agents and block the rest.
Read the entry →x402
HTTP 402 payments for agents
Revive HTTP 402 so an agent can pay per request instead of signing up for an API key.
Read the entry →Crawler access, licensing & indexing
7 termsThe files and headers that tell AI crawlers what they may fetch, what they may train on, what it costs, and when something changed. This is where most sites accidentally block themselves out of AI answers.
llms.txt
/llms.txt
A curated markdown index that points LLMs at the pages you actually want them to read.
Read the entry →llms-full.txt
/llms-full.txt
Your whole documentation set as one clean markdown file an LLM can ingest in a single fetch.
Read the entry →AI Crawler Control in robots.txt
GPTBot, ClaudeBot, Google-Extended and friends
Decide per bot who may crawl, who may train, and who gets nothing.
Read the entry →Content Signals
Content-Signal directives in robots.txt
State whether your content may be used for search, AI input or AI training, separately.
Read the entry →RSL
Really Simple Licensing (/license.xml)
Attach machine-readable licensing and payment terms to your content for AI use.
Read the entry →TDMRep
TDM Reservation Protocol
The W3C-backed way to reserve text and data mining rights under EU copyright law.
Read the entry →IndexNow
IndexNow push indexing
Push URL changes to search and AI indexes instead of waiting to be crawled.
Read the entry →Structured data & machine readability
7 termsHow you make the meaning of a page legible to a machine. These are the oldest techniques in the glossary and the ones that still move the needle most for citation.
JSON-LD Structured Data
Schema.org in JSON-LD
The single highest-leverage way to tell a machine what a page is actually about.
Read the entry →Speakable
schema.org/speakable
Point assistants at the sentences on your page that are worth reading aloud or quoting.
Read the entry →Semantic HTML
Semantic HTML for AI agents
Use the elements that carry meaning so an agent can parse structure without heuristics.
Read the entry →SSR vs CSR
Server-side vs client-side rendering for AI crawlers
Most AI crawlers do not run your JavaScript. If the content is not in the HTML, it does not exist.
Read the entry →Multimodal Markup
Alt text, captions, transcripts and media schema
Give every image, video and diagram a text representation a model can actually cite.
Read the entry →sameAs and Entity Linking
Entity disambiguation via sameAs
Tell the knowledge graph which real-world entity you are, so models stop confusing you with someone else.
Read the entry →Canonical and hreflang
Canonical URLs and language targeting for AI
Make sure the version of a page an AI cites is the one you want to be cited.
Read the entry →No terms match that search. Try a protocol name such as WebMCP, or a file name such as llms.txt.
Stop guessing which of these you are missing
Glippy runs 240+ checks across 16 categories on any page and tells you exactly which of these standards you have implemented, and which you have not.
Every entry was reviewed against its primary sources on . These standards move quickly, so check the linked specs before you ship.