Advertising Compliance Glossary for AI Agents via MCP

Quick Answer

An advertising and platform-policy glossary is a structured reference of the terms used in advertising compliance and platform enforcement — words like shadowban, disclosure, account suspension, DSA, GDPR and FTC endorsement guides — and giving an AI agent access to one makes its compliance reasoning more accurate and consistent. The AuditSocials Glossary MCP server exposes 212 such terms over the Model Context Protocol, so an agent can define a term, search by keyword or category, and list categories, receiving the full definition, applicable platforms, related terms and a persistent URI for each entry. Because the glossary is bundled with the package, it needs no API key and no configuration; it runs over stdio with clients such as Claude, Cursor and VS Code. The data is published under the CC BY 4.0 license, which makes it citable and reusable. The result is that an agent looks up an authoritative definition instead of improvising one, which reduces the risk of confident but wrong compliance statements. Browse the terms on the compliance glossary, add the server with the MCP setup guide, and compare platform rules on the platform comparison.

Predict ad disapprovals before you submit — try the free Meta Rejection Predictor

An Advertising and Platform-Policy Glossary for AI Agents: 212 Terms via MCP in 2026

Why AI Agents Need a Compliance Glossary

Advertising compliance has a dense, specialized vocabulary. Terms like shadowban, demonetization, account suspension, material connection, endorsement disclosure, DSA, GDPR and FTC carry precise meanings, and those meanings differ by platform and jurisdiction. When a person is unsure what a term means, they look it up. When an AI agent is unsure, it tends to do something more dangerous: it produces a fluent, confident definition that may be subtly or completely wrong. In a compliance context, a confident wrong answer is worse than an admitted gap, because it can send an advertiser in the wrong direction.

Giving an agent a reliable reference to consult changes that behavior. Instead of improvising a definition from patterns in its training data, the agent can retrieve an authoritative entry — the actual definition, the platforms it applies to, related terms, and a stable identifier it can cite. This is the same reason human analysts keep a glossary at hand: shared, precise definitions prevent the small misunderstandings that compound into bad decisions. For an agent operating at speed and volume, that grounding matters even more.

"Each entry returns the full definition, category, applicable platforms, related terms and a persistent URI — so a definition can be looked up rather than improvised.
— AuditSocials Glossary documentation"

This guide describes a 212-term advertising and platform-policy glossary built for exactly this purpose, how an agent queries it over the Model Context Protocol, how the definitions are used in practice, and why provenance and licensing make it trustworthy to cite. You can browse the same terms as a human on the compliance glossary, and see how platform rules differ on the platform comparison.

What the Glossary Covers

The glossary is a structured set of 212 terms drawn from the world of social media advertising compliance and platform enforcement. Rather than a flat list of words, each entry is a small record: a definition written to be self-contained, a category, the platforms the term applies to, related terms, and a persistent identifier. That structure is what lets both humans and machines use it precisely — a term is not just defined, it is placed in context.

The coverage spans platform-specific enforcement concepts, cross-platform advertising rules, and the regulatory frameworks that sit above them. A single agent question about, say, whether a piece of content needs a disclosure might touch several of these at once, which is why having them defined consistently in one place is valuable.

The Kinds of Terms Included

Area	Example concepts	Why it matters to compliance
Platform enforcement	Shadowban, demonetization, account suspension, appeal	These describe the actions platforms take and how to respond
Advertising rules	Prohibited claims, restricted categories, ad review	The rules that determine whether an ad runs or is disapproved
Disclosure	Material connection, sponsored content, AI-generated media labeling	What must be revealed to audiences and when
Regulatory frameworks	DSA, GDPR, FTC endorsement guides, COPPA	The laws and guidelines advertising must comply with

Because the terms are cross-referenced through related-term links and tagged with the platforms they apply to, the glossary works as a small knowledge graph rather than a dictionary. That makes it useful for the kind of connected reasoning compliance questions require — moving from a term to its neighbors and to the platforms it touches. For the deeper platform rules behind many of these terms, the Meta ad policies guide is one worked example.

The self-contained style of each definition is a deliberate choice for machine use. A definition that assumes the reader already understands three other terms works poorly when an agent retrieves it in isolation, because the agent may pull a single entry to answer a single question and never see the surrounding context a human browsing the page would. Writing each entry to stand on its own means a retrieved definition is usable as-is, without a chain of further lookups, which keeps the agent's answer both accurate and concise. Where a concept genuinely depends on another, the related-term links make that dependency explicit rather than implicit, so the agent can follow it deliberately instead of guessing at it.

Querying the Glossary Over MCP

The glossary is exposed to AI agents through the Model Context Protocol, an open standard for connecting AI applications to external tools and data. The AuditSocials Glossary MCP server provides three tools, and together they cover the ways an agent needs to interact with a reference: define a specific term, search across terms, and browse the category structure. Because the tools are self-describing, an agent discovers what they do and calls them when a definition would help, without special prompting.

The Three Tools

Tool	What it does	When an agent uses it
glossary_define	Defines a term by name or slug; returns definition, category, applicable platforms, related terms and a persistent URI	The agent knows the term and needs its precise meaning
glossary_search	Searches by keyword and/or category; returns matching terms with short definitions	The agent has a concept but not the exact term
glossary_list_categories	Lists glossary categories with term counts	The agent wants to browse or scope a search by area

The important design choice is that the glossary is bundled with the package, so these tools need no API key and no configuration — the 212 terms ship with the server. That makes it a low-friction addition: an agent gains a citable reference simply by having the server installed. The full setup lives in the MCP setup guide, alongside the compliance-check server that covers the enforcement side.

How Agents Use Definitions in Practice

The value of a glossary tool shows up in the difference between two agent behaviors. Without it, an agent asked to explain whether a creator's post needs a disclosure will generate an answer from memory, which may conflate similar concepts or state a rule with more certainty than is warranted. With the glossary tool, the agent can define material connection, check which platforms the term applies to, follow related terms to sponsored content and endorsement disclosure, and ground its explanation in retrieved definitions rather than recollection.

This retrieval-first pattern is particularly useful for compliance because the field rewards precision and punishes confident errors. An agent that looks up a definition and cites its persistent URI produces an answer a human can verify, which is exactly what you want when the stakes include a rejected ad or a restricted account. The glossary does not replace judgment; it makes the agent's inputs to that judgment accurate.

Where a Glossary Lookup Helps

Explaining a policy term: The agent defines the term precisely instead of paraphrasing it loosely.
Disambiguating similar concepts: Related-term links separate look-alikes such as demonetization and account suspension.
Scoping by platform: Applicable-platform tags tell the agent where a term actually applies.
Producing citable output: A persistent URI per term lets the agent reference a stable source.

Used alongside a live compliance check, the two form a natural pairing: the glossary explains the concepts, and the check applies them to specific content. For how the check side works in an agent workflow, see the pre-publish compliance approach documented across the Compliance API, and track how the rules behind the definitions evolve on the Policy Change Tracker.

There is also an answer-engine dimension to grounding an agent this way. As more questions are answered by AI systems rather than a list of links, the definitions an agent reaches for become the substance of what a user actually reads, and a wrong or vague definition propagates directly into the answer with no intermediate page for the user to sanity-check. An agent that retrieves a precise, sourced definition and attributes it produces an answer that holds up to scrutiny, which is increasingly what distinguishes a trustworthy compliance assistant from one that merely sounds confident. The same structured entries that help an agent also help a person searching for a definition, because a clear, self-contained explanation with its applicable platforms and a stable link is exactly the kind of content answer engines tend to surface and cite.

Provenance, Licensing and Trust

A reference is only as useful as it is trustworthy, and two properties determine whether an agent's citation of a glossary term can be relied upon: where the definitions come from, and whether they can be reused and cited openly. The AuditSocials glossary addresses both. Its definitions are maintained as a deliberate compliance reference rather than scraped text, and the data is published under the Creative Commons Attribution 4.0 (CC BY 4.0) license, which permits reuse with attribution.

Open licensing matters more for AI use than it first appears. When a glossary is CC BY 4.0, an agent can quote a definition and attribute it to a stable source, and downstream products can incorporate the terms without licensing ambiguity. The persistent URI attached to each entry reinforces this: it gives every term a durable address, so a citation remains valid over time rather than pointing at a page that might move. Together, open licensing and persistent identifiers turn definitions into references an agent can safely stand behind.

What Makes the Definitions Citable

Property	What it provides
CC BY 4.0 license	Reuse and citation with attribution, without licensing ambiguity
Persistent URI per term	A durable address so citations stay valid over time
Maintained as a reference	Definitions curated for compliance rather than scraped
Applicable-platform tags	Context that scopes each term to where it applies

This is consistent with a broader principle in compliance content: claims and definitions should be grounded in a source that can be checked, not asserted. The glossary is designed so that an agent citing a term is citing something real and stable. Browse the human-readable version and its sourcing on the compliance glossary.

The same discipline that makes definitions citable also makes them maintainable. Because each term has a stable identifier and a defined scope, the glossary can be corrected and extended over time without breaking the references that point to it — a term's meaning can be refined while its address stays constant. That matters in a field where the underlying rules move: platform policies and regulatory frameworks change, and a reference that can be updated in place, rather than rebuilt, stays accurate without invalidating the citations an agent has already produced. A glossary treated as living infrastructure, not a one-time publication, is what keeps an agent's grounding current instead of slowly drifting out of date.

Adding the Glossary MCP Server

Because the glossary is bundled and requires no key, adding it is the simplest kind of MCP setup: register the server in your client and restart so it discovers the three tools. The server speaks MCP over stdio, so the same configuration shape works across clients such as Claude Desktop, Claude Code, Cursor, VS Code, Windsurf and Cline — you point them at the package command, with no environment variables to set.

Setup at a Glance

Register the server: Add an entry that runs npx -y auditsocials-mcp under your client's MCP configuration.
No key needed: The 212-term glossary ships with the package, so there is nothing to authenticate.
Restart and ask: Restart the client, then ask the agent to define or search a term — for example, "Define account suspension using the AuditSocials glossary."

Once added, the glossary sits alongside the compliance-check server as the reference half of a pair: one explains the terms, the other enforces the rules. Teams building on top of AI assistants can add both to give an agent grounded definitions and a live policy check. The step-by-step configuration for each client is in the MCP setup guide, and the terms themselves are browsable on the compliance glossary.

Checklist for Using the Glossary

[ ] Registered the glossary MCP server in each client that answers compliance questions
[ ] Confirmed no API key is required — the 212 terms ship bundled
[ ] Restarted the client so it discovers define, search and list-categories tools
[ ] Prompted the agent to look up terms rather than define them from memory
[ ] Used related-term links to disambiguate similar concepts
[ ] Scoped terms by their applicable platforms rather than assuming they apply everywhere
[ ] Cited the persistent URI when an answer references a definition
[ ] Attributed reused definitions per the CC BY 4.0 license
[ ] Paired the glossary with a live compliance check for content-specific decisions
[ ] Tracked how the rules behind the definitions change over time

Frequently Asked Questions

Why does an AI agent give better compliance answers with a glossary than without one?

An AI agent gives better compliance answers with a glossary because it can retrieve an authoritative definition instead of improvising one, and in compliance the difference between the two is the difference between a reliable answer and a confidently wrong one. Advertising compliance has a dense, specialized vocabulary — shadowban, demonetization, account suspension, material connection, endorsement disclosure, DSA, GDPR, FTC — and these terms carry precise meanings that vary by platform and jurisdiction. When an agent is unsure what a term means, its default behavior is to generate a fluent definition from patterns in its training data, and that definition can be subtly or entirely incorrect while sounding authoritative. In most contexts a small inaccuracy is tolerable; in compliance it can send an advertiser in the wrong direction, because a confident wrong answer is acted on as if it were right. A glossary changes the agent's behavior from recollection to retrieval. Instead of paraphrasing from memory, the agent looks up the actual entry: the self-contained definition, the category, the platforms the term applies to, related terms, and a persistent identifier it can cite. This grounding matters for the same reason human analysts keep a glossary at hand — shared, precise definitions prevent the small misunderstandings that compound into bad decisions — but it matters even more for an agent operating at speed and volume, where errors propagate quickly. There is also a verifiability benefit. When an agent cites a retrieved definition and its persistent URI, a human can check the source, which is exactly what you want when the stakes include a rejected ad or a restricted account. The glossary does not replace the agent's judgment; it makes the inputs to that judgment accurate, so the reasoning that follows starts from solid ground rather than a plausible guess. You can browse the same 212 terms a human would use on the compliance glossary, and see how platform rules differ on the platform comparison. The distinction is sharpest on look-alike terms that an unaided model tends to blur, such as demonetization versus account suspension, or a platform label versus a legal disclosure duty, where a small conflation produces advice that sounds reasonable but points in the wrong direction; retrieving the precise entry and its related terms is exactly what keeps those distinctions intact. The organizing principle is that a glossary shifts an agent from improvising definitions to retrieving authoritative ones, which is what compliance accuracy requires.

What terms and areas does the 212-term glossary cover?

The glossary is a structured set of 212 terms drawn from social media advertising compliance and platform enforcement, and its coverage spans three connected areas rather than a single flat topic. The first area is platform enforcement: concepts such as shadowban, demonetization, account suspension and appeal, which describe the actions platforms take against content and accounts and how advertisers and creators respond to them. The second is advertising rules: prohibited claims, restricted categories, ad review and similar concepts that determine whether an ad runs or is disapproved. The third is the regulatory layer that sits above the platforms: frameworks like the DSA, GDPR, the FTC endorsement guides and COPPA, which are the laws and guidelines advertising must comply with regardless of any single platform's policy. Disclosure runs through all three areas as a recurring theme, covering material connection, sponsored content and the labeling of AI-generated media. What makes the glossary more than a dictionary is that each entry is a small structured record, not just a definition. Every term carries a category, the platforms it applies to, related terms, and a persistent identifier. The related-term links and platform tags turn the collection into a small knowledge graph: an agent or a person can move from a term to its neighbors and to the platforms it touches, which is exactly the kind of connected reasoning compliance questions require. A single question about whether a post needs a disclosure might touch a platform enforcement concept, an advertising rule and a regulatory framework at once, and having them defined consistently in one place keeps that reasoning coherent. The definitions are written to be self-contained, so a term stands on its own without requiring the reader to already understand three other terms first. For the deeper platform rules behind many of these concepts, the platform guides provide worked examples — the Meta ad policies guide is one — and the full browsable set is on the compliance glossary. Because the collection is curated as a deliberate reference rather than assembled from scraped text, the definitions are written to a consistent standard and scoped to where they actually apply, which is what lets an agent trust that two related entries use their shared vocabulary the same way rather than in subtly different senses that would undermine connected reasoning. The organizing principle is that the glossary covers platform enforcement, advertising rules and regulatory frameworks as a cross-referenced knowledge graph, not a flat word list.

How does an agent query the glossary, and what does each tool return?

An agent queries the glossary through the Model Context Protocol using three tools, which together cover the ways one needs to interact with a reference: defining a known term, searching for a concept, and browsing the structure. The first tool, glossary_define, takes a term by name or slug and returns the full record — the definition, the category, the platforms the term applies to, related terms, and a persistent URI. An agent uses this when it already knows the term and needs its precise meaning, for example when a user asks what account suspension means or when the agent encounters the term while reasoning and wants to ground its understanding. The second tool, glossary_search, takes a keyword and, optionally, a category, and returns matching terms with short definitions. This is what an agent reaches for when it has a concept but not the exact term — searching for disclosure, say, to find the relevant entries rather than guessing the canonical name. The third tool, glossary_list_categories, lists the glossary's categories with the number of terms in each, which lets an agent browse the landscape or scope a search to a particular area before drilling in. Because these tools are self-describing under MCP, the agent learns what each does from the tool's own description and calls the right one when a definition would help, without needing special prompting to do so. A key practical detail is that the glossary is bundled with the package, so none of these tools require an API key or any configuration; the 212 terms ship with the server itself. That makes the glossary a low-friction addition — installing the server is enough to give an agent a citable reference. The server speaks MCP over stdio, so it works with clients such as Claude, Cursor and VS Code using the same configuration shape. The complete, client-by-client setup is documented in the MCP setup guide, which also covers the companion compliance-check server. The no-key, bundled design is more consequential than it sounds for adoption: because there is nothing to authenticate or configure, the glossary can be added to a client in seconds and carries no per-call cost, which removes the usual friction that stops teams from giving an agent a grounded reference in the first place. The organizing principle is that three self-describing tools — define, search and list-categories — let an agent look up, find and browse the 212 terms with no key required.

Why do the CC BY 4.0 license and persistent URIs matter for AI use?

The CC BY 4.0 license and the persistent URIs matter for AI use because together they make the glossary's definitions safe to cite and durable to reference, which is what turns a definition from a nice-to-have into something an agent can stand behind. A reference is only as useful as it is trustworthy, and trustworthiness here has two components: whether the definitions can be reused and cited openly, and whether a citation will remain valid over time. The license addresses the first, and the persistent identifiers address the second. Creative Commons Attribution 4.0, or CC BY 4.0, permits reuse of the material with attribution. For AI use this is more consequential than it first appears. When a glossary is openly licensed, an agent can quote a definition and attribute it to a stable source without licensing ambiguity, and downstream products can incorporate the terms into their own experiences on clear terms. Compare that with definitions of uncertain provenance or restrictive licensing, where reuse is legally murky and citation is discouraged; those are exactly the conditions under which an agent is more likely to paraphrase loosely and lose the source. Open licensing removes that friction and makes citation the natural path. The persistent URI attached to each term addresses durability. A citation is only useful if the thing it points to stays where it is, and a persistent identifier gives every term a stable address that does not break when pages are reorganized. For an agent that cites a definition, this means the reference remains valid over time rather than degrading into a dead link, which is important because compliance answers are often revisited and audited later. Taken together, open licensing and persistent identifiers turn definitions into references an agent can safely cite and a human can reliably verify — the same principle that underlies good compliance content generally, where claims should be grounded in a checkable source rather than asserted. You can see the human-readable definitions and their sourcing on the compliance glossary, and track how the underlying rules evolve on the Policy Change Tracker. This is the same discipline that separates trustworthy compliance content from content that merely sounds authoritative: a definition an agent can attribute to a stable, openly licensed source can be checked by a human, whereas an unsourced paraphrase cannot, and in a field where a confident error carries real cost, that verifiability is the whole point. The organizing principle is that CC BY 4.0 licensing makes definitions freely citable and persistent URIs keep those citations valid, which is what makes the glossary trustworthy for AI use.

How do I add the glossary MCP server, and how does it pair with a compliance check?

Adding the glossary MCP server is the simplest kind of MCP setup because the glossary is bundled and requires no key: you register the server in your client and restart so it discovers the three tools. The server speaks MCP over stdio, so the same configuration shape works across clients such as Claude Desktop, Claude Code, Cursor, VS Code, Windsurf and Cline. In practice, you add an entry to your client's MCP configuration that runs the package command, npx -y auditsocials-mcp, with no environment variables to set because there is nothing to authenticate — the 212 terms ship with the package. After restarting the client so it lists the new tools, you can ask the agent to define or search a term directly, for example asking it to define account suspension using the AuditSocials glossary, and it will call the appropriate tool. The more interesting part is how the glossary pairs with a live compliance check, because the two are designed to complement each other. The glossary is the reference half: it explains what terms mean, which platforms they apply to, and how they relate to one another, giving an agent grounded definitions to reason with. A compliance check is the enforcement half: it takes a specific piece of content and evaluates it against current platform policy, returning a verdict, the risky phrase, the reason and a rewrite. Used together, an agent can both explain the concepts involved in a compliance question and apply them to actual content — defining material connection and then checking whether a specific caption discloses it, for instance. Teams building on AI assistants often add both servers for this reason: the glossary makes the agent's explanations accurate, and the check makes its content decisions safe. Because the glossary needs no key and the check uses a free-tier key, a team can add both and evaluate the pairing at no cost before committing. The step-by-step configuration for each client, covering both servers, is in the MCP setup guide, and the compliance-check side is described across the Compliance API. Adding both servers also mirrors how a careful human works: you look up what a term means before deciding whether a specific piece of content satisfies it, and separating the reference from the enforcement keeps each one focused, so the glossary stays a clean source of definitions and the check stays a current read on live policy, rather than conflating the two. The organizing principle is that the glossary installs with no key and pairs naturally with a live compliance check — one explains the terms, the other enforces the rules.

Don't miss the next policy change.

Create a free account — track every policy change across 8 platforms, get instant alerts, and access every free compliance tool. Or try our Meta Rejection Predictor first.

Create Free Account

Report Keywords — Run AI Compliance Audit

#Compliance Glossary#MCP#Model Context Protocol#AI Agents#DSA#GDPR#FTC#Advertising Compliance#Platform Policy#RegTech#Knowledge Base#Compliance Guide 2026

Share This Report

Tweet Share

How to Add a Pre-Publish Compliance Check to AI Content Workflows with MCP in 2026

AI assistants now draft and schedule social and ad content. This guide shows how to add a pre-publish compliance check over the Model Context Protocol before anything goes live.

Synthetic Media Enforcement Index Q1 2026 — DSA Transparency Database Findings

Q1 2026 DSA Transparency Database snapshot — 299 million enforcement actions across eight major platforms, with the demoted-content layer, automation rates, and EU30 geographic spread broken out.

Platform Holding Company Structure in 2026: Why It Matters for Compliance Intelligence and Risk Mapping

The corporate structure behind each major platform shapes what advertisers can learn from public filings. Knowing which platforms are publicly traded, where they incorporate, and which regulators they answer to is the foundation of platform-policy intelligence.

Platforms

Industries

Knowledge

Blog

An Advertising and Platform-Policy Glossary for AI Agents: 212 Terms via MCP in 2026