> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.markup.ai/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.markup.ai/_mcp/server.

# How Agents Work

> Markup AI's content agents — specialized checks for brand, clarity, accuracy, and AI visibility that apply your organization's standards automatically.

Markup AI **agents** are specialized content checks and optimizations — brand voice, persona reaction, sentence clarity, terminology, AI-search readiness, and more. Configure your brand voice, personas, terminology, and style guide once in the [Markup AI console](https://console.markup.ai), and every agent applies them automatically — in your AI assistant, your authoring tools, and your own integrations.

## Set up once, enforce everywhere

Agents don't ask writers to remember the rules. Your organization's standards live in the console; agents read them at run time. As soon as you update your standards, every future check automatically uses the latest version.

## The agents

| Agent                                              | What it does                                                | Category          |
| -------------------------------------------------- | ----------------------------------------------------------- | ----------------- |
| [AI Voice](/agents/ai-voice-detector)              | Make AI-assisted drafts sound like you wrote them.          | Content Integrity |
| [Brand Terms](/agents/terminology)                 | Keep every word on the approved list.                       | Brand             |
| [Brand Voice](/agents/brand-voice)                 | Sound like your brand, everywhere.                          | Brand             |
| [Claims](/agents/generic-claims)                   | Catch risky claims before compliance does.                  | Accuracy          |
| [Date & Time Accuracy](/agents/date-time-accuracy) | Catch dates and times that can't be right.                  | Accuracy          |
| [Fluff](/agents/focus-agent)                       | Get to the point — and stay on it.                          | Content Integrity |
| [Freshness](/agents/freshness)                     | Never publish stale facts again.                            | AI Visibility     |
| [Key Takeaways](/agents/key-takeaways)             | Put your key takeaways where readers — and AI — look first. | AI Visibility     |
| [Link Integrity](/agents/link-integrity)           | Every link goes where it says it goes.                      | Accuracy          |
| [Persona](/agents/persona)                         | Read your content through your buyer's eyes.                | Brand             |
| [Q\&A Optimization](/agents/answer-authority)      | Headings that answer the question.                          | AI Visibility     |
| [Sentence Clarity](/agents/sentence-clarity)       | Make every sentence easy to read.                           | Content Integrity |
| [Snippet Readiness](/agents/snippet-readiness)     | Write sections that survive being quoted.                   | AI Visibility     |
| [Source Authority](/agents/source-authority)       | Give every claim a source AI search trusts.                 | AI Visibility     |
| [Style](/agents/style-agent)                       | Your whole style guide, applied in one pass.                | Brand             |

## Risk-based scoring

As your agents check content, each issue they surface is assigned a risk level — **High**,**Medium**,**Low** — based on how much risk it poses to your content quality.

Risk levels give you a **consistent way to prioritize across every agent**. No matter which agent flags an issue, the same High/Medium/Low scale lets you triage at a glance and see what to address first — so you can act on what matters most without weighing each agent's findings separately. If you have limited time, focus on High-risk issues; if you're doing a full pass, work all the way down to Low.

| Risk Level | What it means                                                                                                               |
| ---------- | --------------------------------------------------------------------------------------------------------------------------- |
| High       | Issues that are very likely wrong and could undermine credibility, compliance, or brand standards. **Address these first.** |
| Medium     | Issues that may confuse readers or create inconsistency. Worth reviewing, especially for public-facing content.             |
| Low        | Stylistic refinements that improve polish and readability. Address these when time allows.                                  |

## Put agents to work

#### [Markup AI for Google Docs](/apps/google-docs)

The fastest way to put agents to work: catch and fix issues right inside Google Docs, with the Markup AI Chrome Extension.

#### [Connect your AI assistant](/mcp)

Every agent is a tool on the hosted MCP server — Claude, Cursor, and any MCP client.

#### [Build with the API](/developer/agents)

Run any agent with one API call — copy-paste examples for every agent.