Best MCP for Meta Ads in 2026: 20 Servers Compared and Ranked

AdKit is a hosted MCP server built specifically for running ads, not a generic connector. Your agent talks to AdKit, AdKit talks to Meta through the official Marketing API, and every campaign, ad set, ad, and creative your agent touches is saved as a draft you approve before it goes live. You can loosen that in settings once you trust it, but the default protects your budget.

It pairs that safety with the deepest toolkit here. Beyond creating and managing campaigns, your agent gets a searchable library of 500,000+ real ads, AI creative generation, a swipe file, and a strategy skill. The other servers stop at the raw API. It also runs seven ad platforms from one connection, Meta, Google, TikTok, Reddit, X, LinkedIn, and Microsoft, with Pinterest and Snapchat coming, so one agent covers your whole media mix. AdKit is a reviewed Meta Tech Partner, and setup is hosted in roughly three clicks, with no tokens, code, or terminal. It works with Claude, ChatGPT, Cursor, Windsurf, and any MCP client.

You also get extras the bare servers do not have: a free Meta Ads strategy skill built from a decade of ad buying, and no daily request caps, unlike Meta's official beta.

In 2026 Meta shipped its own official MCP server, plus a CLI, so you can manage campaigns from the AI tools you already use. It can create and edit campaigns, ad sets, and ads through natural language, pull reports, and build product catalogs, with no developer credentials or coding required. As a first-party option, it is the most direct line to the Marketing API you can get.

The limit is in the name: it is Meta only. If you also run Google or TikTok, this does nothing for them. Its feature set is account management: campaigns, reports, product catalogs, and signal health, with no competitive research or creative generation. It is free for eligible advertisers in the open beta, within a daily request limit, and there is no draft or approval step, so a change your agent makes applies straight to your live account. Treat it as promising and official, but narrow.

Adspirer is a focused, hosted MCP server with 178+ tools across Meta, Google, LinkedIn, and TikTok. LinkedIn Ads support with 45 dedicated tools is rare in this set, and the free tier with no credit card is a low-friction way to try it. New campaigns are created paused by default, which is a thoughtful baseline for a tool aimed at marketers.

The catch is the metering. Adspirer charges per tool call: the Plus plan gives you 150 calls a month, and a single campaign launch can burn 20 to 50 of them. That puts heavy agent use in the 3 to 7 launches a month range on the entry paid tier, which climbs quickly for anyone running active tests. There is also no dashboard, no ad library, no competitor research, and no creative generation. It is a clean API wrapper with good coverage, but it stops at the API.

Synter has the widest platform coverage of any ad MCP here: 20+ channels including Meta, Google, TikTok, LinkedIn, Reddit, Microsoft, Amazon, Pinterest, Snapchat, Walmart Connect, Instacart, and more. That is more than AdKit. If your media mix stretches across a dozen niche channels, Synter's breadth is the real draw. It is MIT open-source, runs as a hosted server or local CLI, and has tiered autonomy that auto-executes low-risk actions while routing high-impact ones for approval.

The counterpoint for anyone living in Meta: Synter emerged from stealth in March 2026 and has limited independent validation. Its per-platform tooling is templated and shallow. There is no Meta ad library, no competitor tracking, and no swipe file. The free tier is read-only, and the entry paid plan starts at $299 a month, credit-metered, so the cost grows with use. For Meta-specific depth and draft safety, AdKit is ahead. For someone spreading real spend across 15 channels, Synter's coverage is the strongest argument in this list.

Pipeboard is the developer's pick. It wraps Meta, Google, TikTok, and Snap with 120+ tools and exposes every API parameter, making it the most flexible raw-access option here. New campaigns are created paused by default, so accidental spend is contained. It is open-source, runs as a hosted server or CLI, and integrates natively with n8n, Cursor, and custom Python. Founded in August 2025 by an ex-Google SRE, it is a focused, no-nonsense tool.

The trade-off is clear: this is built for engineers, not marketers. You work with raw Meta API objects, which means you need to know what a campaign structure looks like at the API level before you can do anything useful. Edits to existing campaigns execute live with no gate. Execution limits reset weekly, not monthly, so a single busy session can drain the Pro plan's 500 allowance. There is no ad library, no research, and no creative tooling around it. A great building block for technical teams; harder work for everyone else.

Ryze takes a different angle from the rest of the list: it is not just an MCP for managing campaigns, it is an optimization engine that runs 24/7 on your existing campaigns, adjusting bids and budgets server-side between your agent sessions. Meta Ads and Google Ads write capability is confirmed via MCP, with a per-action approval gate so nothing changes without a green light. For advertisers who want continuous optimization rather than just create-and-launch, that is a distinct value.

The things to watch: the credit system is opaque, Ryze does not publish what one credit costs or how many an action consumes, and every plan caps you by ad spend (entry plan has a $25k/month ceiling). Website marketing mentions TikTok and LinkedIn, but the MCP page confirms only Meta and Google, so do not count on those channels. The optional "Schedule mode" removes the approval gate entirely and runs autonomously on your live account. Founded in 2023 by an ex-McKinsey team, it is the most mature purpose-built optimizer here, but it is not a creative or research tool.

Blend is the strongest pick for ecommerce brands on Shopify or WooCommerce. It pulls your product catalog natively, so your agent can build campaigns from your live inventory without manual asset prep. It covers Meta, Google, TikTok, Microsoft Ads, and Pinterest, and at $29/month for the entry tier it is the lowest entry price among the purpose-built ad MCPs here. Founded in 2020 with $3.6M raised, it is also the most established tool in this section of the list.

The caveats: every tier has an operations cap (500/2,500/10,000 per month), and an active agent burns through those faster than you might expect. Actions execute directly on live accounts with no draft or approval step, so there is no buffer between your agent and real spend. If you are not running an ecommerce brand on Shopify or WooCommerce, Blend was not built with you in mind. For ecommerce in that stack, it is the natural fit. For everyone else, the missing draft layer is the thing to weigh.

Zuckerbot is an open-source (MIT) CLI and MCP server for Meta Ads and an approved Meta partner, with 50 tools covering campaigns, audiences, creatives, and conversion tracking. It has an approval step before a campaign goes live, which is more than some bigger names offer, and the free tier is read-only so you can explore safely. It is the most credible of the open-source servers here.

The trade-offs are where it parts ways with a tool like AdKit. It is Meta only, so it does nothing for Google, TikTok, or LinkedIn, and you install it from the terminal with an API key, which rules out anyone who does not live in a shell. The pricing is the part to watch: paid plans add a percentage of your ad spend on top of the monthly fee ($49/mo plus 0.1% of spend, $149/mo plus 0.05%). The more you spend, the more it costs you, which is the opposite of how a flat tool scales.

The one real reason to choose Zuckerbot over AdKit is that you specifically want open-source code you can self-host. On coverage, setup, the toolkit around it, and how the price behaves as you scale, AdKit is ahead, even if Meta is the only channel you run.

This is an open-source Meta Ads MCP server (39 tools) that runs as a Cloudflare Worker in your own account. It handles campaign management, ad sets, creatives, audiences, reporting, and batch workflows, so it can fully write to your account. For a developer who wants total control and no third party in the middle, that is appealing.

It is the most do-it-yourself option here. It does not ship with a hosted control plane, a shared Meta app, or an end-user dashboard. You deploy it to Cloudflare with your own Meta app credentials, wire up JWT auth, and maintain it yourself. There is no approval gate, so a write goes straight to your live account. It is a building block for engineers, not a product for marketers.

This is an open-source (MIT) server distributed on npm, with a sprawling 135 tools covering campaigns, ad sets, ads, and creatives. It can create, update, pause, and delete across your account, so on raw surface area it does a lot. Of the small open-source repos here, it is one of the more active.

It is still a solo project with no hosted product, no company, and no support. There is no approval or draft gate, so a write goes straight to the Meta Marketing API and your live budget. You run it yourself, in your own environment, with your own Meta credentials. Plenty of tools, none of the guardrails or service around them.

This is an open-source (MIT) TypeScript server with 35 read tools always on and 19 write tools you have to switch on with an environment variable. New entities default to paused. For a small repo, that is a careful design: it is hard to spend money by accident.

The catch is everything around it. It is unofficial and not affiliated with Meta, a single-maintainer project, and you host it and wire up your own credentials. It is the most responsibly built of the obscure repos, but it is still a repo, not a product.

This is an open-source (MIT) server that exposes Facebook Ads as REST-style endpoints with MCP framing. It can create, update, and delete campaigns, ad sets, and ads, so it covers the basics of writing to an account.

It is a starting point, not a product. The repo has a single commit on its main branch, so it was published once and not iterated. There is no approval gate, no hosted option, and no real documentation. You deploy it yourself, often from a Railway template, and own everything after that.

This is a small open-source (MIT) Python server with just five functions: it can create full campaigns, then pause, activate, or delete them. Despite the size, it is thoughtful about safety. New campaigns are created paused, and the create function defaults to a dry run.

But it is unmistakably a weekend project: a handful of commits, no docs site, and no company behind it. Five functions is enough to demo, not to run a real account day to day. Take it as a clean reference, not a tool you would trust with live spend at scale.

This is an open-source (MIT) Python server built on FastMCP that wraps the Facebook Business API. It is the earliest-stage option here, still on version 0.0.9.

It is barely documented, and it is not clear whether it can write to an account or only read, so we mark write support as unknown rather than guess. Still pre-1.0, it is something to watch, not something to run your ads on.

Composio is a hosted connector platform, closer to a Zapier for AI agents than a Meta Ads tool. Meta Ads is one toolkit on a marketplace of 1,000+ app connectors. For Meta it exposes a dedicated 52-tool Meta Ads toolkit (create campaigns, ad sets, ads, audiences, update budgets), so if you are wiring one agent across a CRM, a spreadsheet, and a dozen other tools, having ads as one more connector is handy.

Judged as an ads tool, it is thin. There is no draft or review system, so a write goes straight to your live account. You supply your own Meta OAuth app and scopes, and for real write access that app has to clear Meta's app review and business verification, which is fine for engineers and a wall for everyone else. It is a general connector that happens to reach Meta, not a tool built to run ads.

Funnel.io is an established marketing data hub, and its Funnel MCP lets AI tools like Claude and Cursor read directly from the data you have already centralized, including Meta ad performance pulled through one of its 600+ connectors. If your job is analysis, asking an agent how last week's Meta campaigns did against clean, unified data, it is a credible option.

But it is read-only. The MCP can pull numbers; it cannot create a campaign, change a budget, or launch an ad. So it is not a tool for running Meta ads, it is a tool for reporting on them. It is hosted, sold to teams on a custom plan with no public price, and we did not find a Meta partner badge. Useful next to a management tool like AdKit, not a replacement for one.

Triple Whale is a well-known ecommerce analytics platform, and it ships an open-source MCP server that turns plain-language questions into calls against your Triple Whale data, surfacing profit, revenue, and Meta ad performance side by side. For a Shopify brand that already lives in Triple Whale, reading ad results through an agent is a natural fit, and it is a certified Meta partner.

Like Funnel, it is read-only: it reports on Meta ads, it does not manage them. You self-host the server with a Triple Whale API key and an active subscription, so it only makes sense if you already pay for the platform. A solid analytics companion, not a way to run campaigns.

Windsor.ai is a data integration platform that connects 325+ marketing sources to BI tools and warehouses, and its MCP exposes six read-only tools for pulling that data into AI clients. Founded in 2017 and acquired by team.blue in January 2026, it is the most mature and stable company in this list, with SOC 2 Type II certification, GDPR compliance, and 7,000+ customers across 120 countries.

Like Funnel and Triple Whale, this reads Meta data, it does not run ads. The 6 MCP tools are very shallow compared to a management server, and Meta is one source among hundreds. Row overage billing and steep jumps between tiers add cost complexity. The enterprise trust posture is the strongest of any read-only tool here, but the use case is data integration, not campaign management.

Adzviser is a cross-channel reporting connector that pipes data from 45+ sources into AI clients, spreadsheets, and BI tools. Every plan includes all 12 output destinations (Google Sheets, Excel, Claude, Gemini, Looker Studio, Power BI, BigQuery, and more), and the Google Sheets add-on has 2,000+ installs. Founded in 2022 by an ex-Google Ads engineer, it is a legitimate established tool in the reporting space.

The framing is important: this reports on Meta ads, it does not run them. There is no write capability of any kind. The tier structure is also steep, jumping from 3 sources at $48.99 to 7 sources at $148.99 with nothing in between. Some user reviews flag accuracy gaps versus native platform dashboards. A useful reporting companion if spreadsheets and BI are central to your workflow, not a substitute for a management tool.

Flyweel is an "Ad Spend OS" with a read-only analytics MCP covering 222 metrics across Google (132), Meta (67), and TikTok (23). The MCP is free with unlimited tool calls on both tiers, which is the most generous free analytics offering in this list. Its standout feature is CRM-to-ad-spend attribution across 9 CRMs (HubSpot, Salesforce, Pipedrive, and more) plus Xero invoice reconciliation. For teams that want to close the loop between ad spend and closed revenue in their AI client, few analytics MCPs do that.

The constraint is the same as all the tools in this section: Flyweel reports on Meta ads, it does not run them. Eight MCP tools, only one of which does actual analytics, and data runs 1 to 3 hours behind real-time on Meta. Founded in March 2025 with Antler pre-seed backing, it is very new. But the free unlimited MCP and the attribution angle make it the most interesting read-only option for teams with a finance or ops lens.