Skip to main content
The AdAdvisor MCP server gives your AI assistant (Claude, ChatGPT, Cursor, Codex, etc.) direct access to your Meta (Facebook and Instagram) ad data. Ask questions in natural language, get answers grounded in your actual campaign performance. But it goes beyond raw data. Because AdAdvisor knows your business context (your break-even ROAS, average order value, target CPL, budget, and more), the AI isn’t just pulling numbers. It can evaluate whether a 5x ROAS is good or bad for your business. It knows your goals, your margins, and your landing page. That context makes the difference between generic data retrieval and genuinely useful analysis.

Why use this instead of the Meta API directly?

If you’ve ever tried to build your own Meta Ads integration, you know the pain:
  • Creating and configuring a Meta developer app
  • Managing OAuth tokens, refresh flows, and permissions
  • Dealing with rate limits and API versioning
  • Figuring out which fields to request and how to structure queries
AdAdvisor handles all of that for you. You connect your Meta account once through a simple OAuth flow, and we take care of authentication, token management, and data syncing behind the scenes. On top of that, the MCP tools are designed specifically for LLMs. The responses are structured and annotated so your AI assistant gets the most useful context possible, not just raw API dumps.

How to get started

The full setup takes about 5 minutes.
1

Create your AdAdvisor account

Sign up at app.adadvisor.ai. You get a 7-day free trial with full access to all features, including the MCP server.
Even if you cancel after the trial, you keep access to the MCP server. It’s included at every plan level.
2

Set up your organization and business

During onboarding, you’ll create an organization (your team workspace) and a business. For the business, you’ll enter key metrics that help the AI understand your goals:
  • Average order value (AOV)
  • Break-even ROAS
  • Target cost per lead (CPL)
  • Monthly budget
  • Other profitability metrics relevant to your business
These aren’t sensitive financial details. They’re the benchmarks the AI uses to evaluate your ad performance in context.
3

Add your website or landing page

Enter your website URL and AdAdvisor will crawl it. This generates a business profile and insights that the MCP server exposes as a resource. When your AI assistant queries your data, it has context about what you sell, how you position it, and what your landing experience looks like.
4

Connect your Meta ad account

Link your Meta account through a secure OAuth flow. AdAdvisor will start syncing your campaign, ad set, and ad data automatically. No developer app, no API keys, no token management on your end.
5

Wait for the initial sync

The first sync pulls in your historical data. This usually takes a few minutes depending on account size. After that, data stays up to date automatically.
6

Connect the MCP server to your AI tool

Once your data is synced, set up the MCP server in your preferred AI client. We support Claude Code, Claude Desktop, ChatGPT, Cursor, VS Code, Gemini CLI, Codex, Windsurf, and any other MCP-compatible client.

What the MCP server exposes

Once connected, your AI assistant gets a full toolkit for working with your Meta ad accounts. This goes well beyond reading data. Your AI can build campaigns, update targeting, manage creatives, and adjust budgets, all through natural conversation.

Read your account

Browse everything in your ad account: campaigns, ad sets, ads, creatives, and custom audiences. Ask questions like “show me all my active campaigns” or “what ad sets are running in my retargeting campaign?”.

Performance and analytics

Pull aggregated KPIs (spend, revenue, ROAS, CPA, CPM, CTR, conversions) across any date range, see daily time-series trends for a specific campaign, and check your pixel health to catch tracking issues before they cost you money.

Targeting and audiences

Search Meta’s full targeting catalog (interests, geo, behaviors, job titles, life events, and more), estimate audience sizes before you commit, and create website custom audiences and lookalikes from conversation. You can also update the targeting on existing ad sets, with validation against Meta’s reach estimate before changes apply.

Creative management

Browse the images and videos in your ad account through an interactive visual picker (click to select, no copy-pasting hashes or IDs), preview existing creatives with their engagement metrics so you can reuse ones with social proof, upload new assets through a drag-and-drop widget, and create new ad creatives from images, videos, or existing Facebook and Instagram posts. Multi-variant copy is supported (up to 5 headlines, messages, and descriptions per creative).

Campaign creation

Build a full campaign from scratch: campaign, ad sets, and ads. Everything gets created paused so you can review before going live.

Optimization and control

Pause or resume any campaign, ad set, or ad, and adjust daily or lifetime budgets. Changes take effect on Meta immediately.

Duplication

Copy campaigns, ad sets, or ads to scale what’s working. Duplicates are created paused so you can adjust before activating.

Business context (automatic)

Every tool call is grounded in your business context. The MCP exposes your break-even ROAS, AOV, target CPL, daily budget cap, account currency and timezone, and the AI-generated business report from crawling your website. Your AI isn’t just pulling numbers. It can tell you whether a 5x ROAS is good or bad for your business, which campaigns are underperforming relative to your targets, and what to change.
Tools that create or change things always ask for your confirmation first. Nothing happens on your ad account without you saying yes.
For the full list of every tool with example prompts, see Available Tools.

What’s coming next

We’re actively expanding what the MCP server can do:
  • More advertising platforms. Google Ads, TikTok, and others, alongside Meta.
  • Deeper analytics. Attribution modeling, cohort analysis, and automated reporting.
  • Deeper AdAdvisor integration. Trigger recommendation runs, review and apply AI suggestions, and manage your account, all from your AI tool.

Want to learn more about AdAdvisor?

The MCP server is one part of the platform. AdAdvisor also includes a full ads manager, AI-powered campaign recommendations, a campaign wizard for building new campaigns from scratch, and more.

What is AdAdvisor?

Learn about the full platform and everything it can do for your ad performance.

Set Up Your MCP Client

Ready to connect? Follow the setup guide for your specific AI tool.

Create API Keys

Generate keys if you prefer API key auth over OAuth.

What is the MCP Server?

Technical details on how authentication and data access work.
Last modified on April 11, 2026