For AI coding agents

Turn your AI agent into a content engineering machine.

Ground your agent in continuous AI visibility monitoring. Arm it with your brand voice, knowledge base, and a pipeline of content ideas that close real gaps in AI conversations.

Add Pendium as a custom connector in your AI agent

https://pendium.ai/api/mcp

Create account or sign in

Then add the connector above and authorize with your account when your agent prompts you. No API key needed.

// PENDIUM — AI Visibility + Content Engineering MCP
//
// Pendium tracks how ChatGPT, Claude, Gemini, Perplexity, Grok, DeepSeek, and
// Google AI Overviews perceive, mention, and recommend brands — then helps you
// close the gaps. This one server gives your agent 80+ tools across six areas,
// so it can go from "how do AI platforms see us?" all the way to drafting and
// publishing the content that changes the answer.
//
// WHAT YOUR AGENT CAN DO:
//
//   VISIBILITY        See how every AI platform ranks and describes you.
//     scan_visibility, get_scan_status, get_report, get_recommendations,
//     complete_action, get_scan_history, get_cited_sources,
//     lookup_brand_score (free, no auth), trigger/get_visibility_scan_preview
//
//   CONTENT STRATEGY  The brand context every good draft needs.
//     get_goals, create_goal, get_brand_voice, get_factsheet,
//     get_company_profile, analyze_url
//
//   CONTENT WORKFLOWS Turn a gap into finished, published posts.
//     create_workflow, generate_workflow_ideas, run_workflow_pipeline,
//     list_blog_posts, regenerate_blog_post, publish_blog_post, schedule_blog_post
//
//   AGENT SITE        Build and publish pages on the brand's hosted site.
//     create_site_page, generate_site_page, publish_site_page,
//     generate_service_area_pages, set_site_public
//
//   KNOWLEDGE         Ground writing in real brand facts.
//     query_knowledge, list_knowledge_sources
//
//   CONFIGURATION     Tune the scans and manage the account.
//     get_account, add_visibility_topic, add_visibility_persona, set_language
//
// Full tool reference: https://pendium.ai/docs/tools
// Every response includes a _pendium field with context and next steps, so your
// agent can discover the rest of the tools and work out the workflow itself.
//
// SETUP — works with Claude Code, Cursor, Windsurf, and any MCP-compatible agent.
// Authentication is handled automatically via OAuth — you'll sign in with your
// Pendium account when you first connect.
//
// Add the JSON below to your agent's MCP config file:
//   Cursor:      ~/.cursor/mcp.json
//   Claude Code: ~/.claude.json (or run: claude mcp add pendium)
//   Windsurf:    ~/.codeium/windsurf/mcp_config.json
//   Claude Desktop: ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "pendium": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://pendium.ai/api/mcp"]
    }
  }
}

// After adding, restart your agent (reload window in Cursor, restart CLI for Claude Code).
//
// INSTALL WORKFLOW SKILL (optional — teaches your agent Pendium best practices):
//   Claude Code:  mkdir -p ~/.claude/skills/pendium && curl -sL https://pendium.ai/skills/pendium/SKILL.md -o ~/.claude/skills/pendium/SKILL.md
//   Cursor:       mkdir -p ~/.cursor/skills/pendium && curl -sL https://pendium.ai/skills/pendium/SKILL.md -o ~/.cursor/skills/pendium/SKILL.md
//
// Then ask: "Scan AI visibility for https://your-website.com and show me the gaps"

See live examples

What your agent gets

Everything it needs to get you recommended.

Visibility Scores

Your agent sees how every AI platform perceives you

Detailed scores across ChatGPT, Claude, Gemini, Grok, Perplexity, DeepSeek, and AI Overviews — broken down by customer persona, topic, and query. Your agent knows exactly where you're strong, where you're weak, and where you're invisible.

Content Ideas

Get a backlog of content ideas grounded in real data

Your agent pulls strategic content goals and specific content pitches generated from your visibility scan. Each idea targets a real gap in AI conversations — not keyword guesses. Unused ideas are prioritized so your agent always knows what to write next.

Brand Voice

Write on-brand content every time, automatically

Your agent gets your full brand voice — tone traits, style guide, writing examples, and company profile. When it creates content from visibility insights, it writes in your voice with your positioning, not generic AI copy.

Competitor Intel

See who AI recommends instead of you

For every query your customers ask, your agent sees which competitors get mentioned, how often, and on which platforms. It uses this to identify what they're doing differently and where you can overtake them.

Action Items

Prioritized actions your agent can execute on

Scan recommendations become prioritized action items — content to create, sources to reach out to, technical improvements to make. Your agent filters by priority and status, works through the list, and marks items complete as it goes.

Knowledge Base

Ground every piece of content in real brand facts

Your agent searches semantically across your brand's knowledge base — website content, uploaded documents, and training materials. When it writes, it pulls real claims, product details, and positioning language so nothing is hallucinated.

Score Trends

Measure the impact of every piece of content

Historical scan data shows how your AI visibility changes over time. Your agent tracks whether content is moving the needle, which platforms are responding, and whether competitors are gaining or losing ground.

How it works

From connector to content, in one loop.

Add the connector to your agent

Pick your agent, copy the snippet, and add Pendium's MCP server as a custom connector. Sign in once with your account to authorize — your agent now has access to Pendium's visibility API. The whole setup takes ten seconds.

Ask about your brand in plain English

"How do AI platforms see my brand?" — your agent calls Pendium, gets structured visibility data, and reports back in context. No API docs to read, no syntax to learn.

Pendium scans across every major AI platform

Real buyer personas run real queries across ChatGPT, Claude, Gemini, and AI Overviews. Your agent gets back scores, gaps, competitor mentions, and cited sources — everything it needs to act.

Your agent identifies what to create

Using the visibility data, your agent pinpoints the topics and angles where you're missing from AI conversations. It suggests content that targets real openings — not keyword guesses.

Create and publish content that closes the gaps

Your agent drafts articles, guides, and comparisons grounded in real visibility intelligence — then schedules and publishes them, either to your Pendium-hosted site or straight into the tools you already use. Content designed to help AI platforms understand your business and recommend you.

Monitor continuously, improve continuously

Set up recurring scans so your agent always has fresh data. When visibility shifts — a competitor gains ground, a platform changes behavior — your agent catches it early and adapts.

Agent skill

Drop the Pendium skill into your coding agent.

For Claude Code and Cursor, install the Pendium skill so your agent always knows how to run a full content-engineering workflow — scan, analyze, and create — without prompting.

Claude Code

~/.claude/skills/pendium/SKILL.md

Cursor

~/.cursor/skills/pendium/SKILL.md

Close the loop

Give your agent AEO superpowers.

Without visibility data, your agent is writing into the void. It doesn’t know where AI platforms are sending people instead of to you, which topics you’re invisible on, or whether last month’s content moved the needle at all.

Pendium gives your agent a continuous feed of what’s working, what’s not, and where the gaps are — so every piece of content it creates is aimed at a real opening in the conversation, not a best guess.

Without Pendium

Agent writes content based on keyword research and best guesses
No idea what AI platforms actually say about you
Content may be great but invisible to AI agents
No brand voice consistency, no knowledge grounding
Reactive — you find out you're missing after the fact

With Pendium

Agent writes content grounded in real AI visibility data
Knows exactly where you're strong, weak, and missing
Content targets the specific gaps in AI conversations
Writes in your brand voice, grounded in your knowledge base
Proactive — continuous monitoring and trend tracking catches shifts early

Continuous monitoring

Know the moment a competitor overtakes you

Set up recurring scans and your agent gets a continuous feed of how AI platforms perceive your brand. When visibility shifts — a competitor gains ground, a new platform emerges — your agent catches it and adapts your content strategy automatically.

Topics and queries tracked across AI platforms

Real buyer personas

Define custom personas that map to your real customer archetypes

A price-conscious shopper asks differently than an enterprise buyer. Pendium simulates real customer personas and runs their actual queries — so your agent sees how AI responds to every type of buyer in your category, not just generic searches.

Persona cards showing different customer types

Data-grounded content

Publish content that actually moves the needle

Your agent doesn't guess what to write. It uses visibility data to identify the specific topics, angles, and platforms where you're missing from AI conversations — then creates content designed to close those gaps and get you recommended.

Content dashboard showing AI-optimized articles

FAQ

Questions, answered.

What is this, exactly?

An MCP server. MCP (Model Context Protocol) is the open standard that lets an AI agent connect to an outside tool. When you add Pendium's URL to Claude Code, Codex, Cursor, or another agent, your agent gains a set of Pendium tools it can call on its own — pull your AI-visibility scores, read your brand voice, draft a post, publish a page. You keep talking to your agent in plain English; it decides which Pendium tools to call and when.

Do I need to learn an API to use this?

No. If you already know how to add an MCP server, the snippets above give you the exact command or config for your agent. If you don't, copy the plain-English prompt, paste it into your agent, and it will install the connector and walk you through signing in. Either way, after that you just talk to your agent — it handles the tool calls behind the scenes.

Which agents does this work with?

Claude Code, Codex, Cursor, Windsurf, Claude.ai, ChatGPT, and any MCP-compatible agent. They all connect the same way: add the MCP server URL and sign in with your Pendium account over OAuth when your agent prompts you. There's a tab above with the precise setup for each.

Do I need an API key to start?

No. MCP connections authenticate over OAuth — you sign in with your Pendium account and your agent gets access automatically. API keys are only for hitting the REST API directly (curl, scripts, no-code tools), and you can create one from your dashboard whenever you need it.

What can my agent actually do with Pendium?

More than 80 tools across six areas: visibility monitoring (scans, reports, trends, competitor mentions, cited sources, recommendations), content strategy (goals, brand voice, factsheet, company profile), content workflows (generate, draft, schedule, and publish blog posts), site (create and publish pages on your hosted brand site), knowledge base (semantic search over your own content), and configuration (topics, personas, account). It's a full content-engineering loop, not just a read-only dashboard.

What AI platforms does Pendium monitor?

ChatGPT, Claude, Gemini, Grok, Perplexity, DeepSeek, and Google AI Overviews — the major platforms consumers and businesses use to research products and services.

How is this different from just asking my agent to search the web?

Web search tells you what's on the internet. Pendium tells you what AI platforms actually say when someone asks about your category — which brands they recommend, which they ignore, and why. That's the data your agent needs to create content that changes the answer.

Does Pendium create the content too, or just the strategy?

Both, and you choose. The MCP can draft, generate images for, schedule, and publish blog posts and agent-site pages directly through Pendium — grounded in your visibility data, brand voice, and knowledge base. Or, if you'd rather write in a system you already use, your agent can pull the same data and strategy and create the content there. Pendium is the eyes and the strategy; it can also be the hands when you want it to be.

How does the brand voice feature work?

When you onboard a brand, Pendium analyzes your website and builds a brand voice profile — tone traits, style guide, writing examples, and company positioning. Your agent pulls this before writing content, so everything it creates matches your brand's voice and personality.

Start in 30 seconds.

Add the connector, sign in with your account, and ask your agent about your brand. Your agent handles the rest.

Get the snippet Book a demo