Meta Ads Engine

Meta Ads Engine

A clean-room Meta Ads MCP server that lets you build, launch, and read Meta ad campaigns via Claude Code using a single system-user token.

Category
Visit Server

README

Meta Ads Engine

A clean-room Meta Ads MCP server plus a battle-tested media-buying playbook. It runs the front end of a Meta ads operation inside Claude Code: read the account, build campaigns, upload video creatives, launch a paused three-ad-set structure, and read performance. One Meta system-user token, no third-party auth broker.

New here? Read START-HERE.md first for setup. The method lives in PLAYBOOK.md.

What it is

  • Our own MCP. The server is written from scratch against the Meta Graph API (Python, FastMCP + httpx). No forked code, no vendored package, MIT licensed. It talks straight to Graph with your system-user token.
  • Our own playbook. PLAYBOOK.md is the Jeremy Haynes post-Andromeda method plus the exact creative-upload flow that works around Meta's DCO block. The tool code encodes the hard-won gotchas so you do not have to remember them.

What it does

  • Read. List ad accounts and pages, pull campaigns and ad sets and ads, read insights (spend, reach, ctr, cpc, cpm, actions).
  • Build. Create a paused campaign, three ABO ad sets (broad, interest stack, lookalike stack), multi-variant creatives, and ads.
  • Video creatives. Upload a local video to the ad account and build a creative from it, thumbnail handled.
  • Launch. The /launch skill drives the whole build to a paused, live-ready campaign. It never sets anything live without your yes.

What it does not do (honest)

  • It does not optimise live campaigns for you. The scaling loop (duplicate, keep the no-reach ads, pull the winners) is a plan you follow, not an automation.
  • It does not generate creative. It uploads and structures what you give it.
  • It does not set campaigns live on its own. Everything ships PAUSED.

Tools

Tool What
get_ad_accounts List accessible ad accounts
get_account_pages List promotable Pages on an account
get_campaigns / create_campaign / update_campaign Campaign CRUD (PAUSED by default, ABO)
get_adsets / create_adset / update_adset Ad set CRUD (budget in cents, Advantage+ audience off)
create_ad_creative / get_ad_creatives Multi-variant (asset_feed_spec) or single-variant (placeholder) creatives
create_ad / update_ad / get_ads Ad CRUD, including the DCO-bypass swap
upload_video / get_video_thumbnail Ad-account video upload + fresh thumbnail
get_insights Performance read at account/campaign/adset/ad level
search_interests Interest search for the interest-stack ad set

The launch flow (how a build runs)

  1. create_campaign (PAUSED, an OUTCOME_* objective).
  2. create_adset x3: broad, interest stack (search_interests to fill it), lookalike stack. ABO budgets in cents.
  3. upload_video per creative, get_video_thumbnail for a fresh thumbnail.
  4. Creative upload uses the DCO bypass: build a single-variant placeholder creative, create_ad against it, then update_ad to swap in the real multi-variant creative. See PLAYBOOK.md for why.
  5. Review in Ads Manager. Set live only on an explicit yes (update_campaign status ACTIVE).

Stack

Python 3.11+, FastMCP, httpx. Runs over stdio as an MCP server. MIT licensed.

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured