Community

Cross-platform ad management MCP server for Google, Meta, TikTok, LinkedIn, Amazon & ChatGPT Ads

How to install ↓ Compare GitHub source repository ↗

Category: Dev Tools #158 of 438
Stars: ★ 99 Popular
Transport: Streamable HTTP
Runtime: Hosted (no local runtime)
Credentials: API key / credential required
License: MIT
Last commit: 3 days ago
Tools: 6

57FMRS · C

A comprehensive cross-platform ad management MCP server supporting major ad platforms with a rich toolset and automation features. However, it is an unofficial third-party service requiring cloud interaction and third-party authorization, so users should evaluate data privacy and costs.

Strongest · Setup experience 14/20 Weakest · Reliability 8/20

Reliability

8/20

Security and permissions

12/20

Maintenance

9/20

Documentation

14/20

Setup experience

14/20

Why each score

Reliability 8/20: Evidence: Manifest declares a remote Streamable HTTP server (https://mcp.adspirer.com/mcp), README describes protocol 2025-03-26 with fallback to 2024-11-05, tool annotations (readOnlyHint/destructiveHint), and a health-check endpoint; repo includes a 62-check offline validation script (scripts/validate.sh) and a CI mode (sync-skills.sh --check). Deductions: static review only; nothing executed; the server's core paths (actual tool implementations) are not in the supplied material; README's 400+ tool claim vs. the manifest's 400+ claim have a minor count inconsistency (README table sums to 392 listed tools); no committed end-to-end tests verifying the MCP handshake or tool behavior were shown. Therefore a mid-range score below the 12 cap.
Security and permissions 12/20: Evidence: OAuth 2.1 authorization code + PKCE (RFC 8252/7591/9728) with dynamic client registration; README explicitly states no conversation logging, only tool requests are processed; REST alternative requires Personal Access Tokens (sk_live_); all tools carry MCP safety annotations; new campaigns are always created paused; SECURITY.md vulnerability reporting channel exists. Deductions: no security policy/audit evidence; the curl|bash install scripts distributed by the README cannot be statically verified; least-privilege for tools is not confirmable (annotations are claims, not auditable enforcement); REST API token lifecycle details are not disclosed; no direct red-line violation is evident, but the server-side implementation cannot be confirmed, capping the score at 12.
Maintenance 9/20: Evidence: Repository is unarchived, version 1.2.0, 3 open issues, versioned release field, SECURITY.md and support email, and README documents ongoing feature expansion (ChatGPT Ads, plugins, Claude Tag submission). Deductions: license is proprietary/unknown — README explicitly says 'Proprietary' and the manifest has no license field; no commit/release history or issue-response evidence was provided; no evidence of dependency updates or security fix cadence; popularity signals (stars, issues) are discovery signals only and add no points.
Documentation 14/20: Evidence: Layered documentation — user-facing README with per-client install tables, concrete config examples for Claude/ChatGPT/Cursor/Codex/Gemini/OpenClaw over MCP, example prompts, platform/tool matrix, REST API reference, CONNECTING.md, PRIVACY.md, TERMS.md, SECURITY.md, SUPPORT.md, plus developer docs for architecture, template sync, adding platforms, adding IDEs, and a documented 62-check validation suite. Deductions: pricing is given but rate limits, per-tool parameters, troubleshooting, and service-status history are not evidenced; tool-count inconsistencies exist; overall strong documentation with minor gaps.
Setup experience 14/20: Evidence: Direct URL configuration is provided: Claude connector, ChatGPT connector, Cursor mcp.json JSON block, Codex config.toml block, gemini extensions install, openclaw plugins install, and a one-line curl installer; OAuth 2.1 is described as completing in the user's browser so the assistant never sees credentials; free tier and priced tiers with clear prices. Deductions: no verifiable execution evidence in this static review (CI workflows or committed tests exercising the install); the curl|bash install scripts cannot be audited from supplied material; setup guidance for less-mainstream clients (e.g., Perplexity/Manus) is generic; no proof of 'try it' troubleshooting. Hence below the 15 cap.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessUses the networkChanges third-party account data

Best for

Ad marketers
Digital marketing agencies
Businesses managing multiple ad accounts and platforms
Teams using AI assistants for ad optimization

Not for

Users who do not need ad management features
Users who want a fully local solution without cloud dependencies
Users with a limited budget who only need simple ad tools

Required permissions

Access your ad platform accounts (via OAuth)
Read and modify campaigns, ad groups, ads, etc.
May create campaigns (but defaults to paused state)
Access account reports and performance data

Risks and side effects

Ad data may be sent to Adspirer servers for processing
Third-party authorization required, potential credential exposure risk (though OAuth is used and passwords are not stored)
Campaigns could be accidentally launched, but default is paused requiring human approval
Free tier has monthly call limits

Setup

Before you start

Runtime：Hosted (no local runtime)

In a supported client (e.g., ChatGPT), install the Adspirer app or plugin, or add a custom connector with URL https://mcp.adspirer.com/mcp.
Complete OAuth 2.1 login (Google or email).
Connect your ad platform accounts at adspirer.ai/connections.
Verify the connection with a natural language prompt like 'Check my connected ad platforms'.

claude_desktop_config.json

{
  "mcpServers": {
    "adspirer": {
      "url": "https://mcp.adspirer.com/mcp"
    }
  }
}

Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json

{
  "servers": {
    "adspirer": {
      "url": "https://mcp.adspirer.com/mcp"
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal

claude mcp add --transport http adspirer https://mcp.adspirer.com/mcp

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

After connecting your client to https://mcp.adspirer.com/mcp and completing the OAuth 2.1 sign-in, run /mcp in Claude Code to confirm the adspirer server appears, or ask your assistant "Check my connected ad platforms" — a successful answer proves the connection.

Troubleshooting

Check MCP connection URL correctness
Verify OAuth login and authorization completion
Check ad platform connections are valid (at adspirer.ai/connections)
Check server status at mcp.adspirer.com/health
Check if free tier call limit is reached

Things to try

Once connected, you can ask your AI assistant things like:

Research keywords for my emergency plumbing business in Chicago and show high-intent keywords with real CPC data.
Show me campaign performance for the last 30 days across all platforms and tell me what to optimize.
Create a Google Performance Max campaign for luxury watches targeting New York with a $50/day budget.
Research keywords for Google Ads and create a LinkedIn sponsored content campaign targeting small business owners for my handmade jewelry business.

Tools 6

campaign_creation writes

Create ad campaigns: Google Search, Performance Max, Display, Demand Gen, YouTube, Meta image/video/carousel/lead-form, LinkedIn sponsored content/lead-gen, TikTok in-feed/Spark/carousel/app promotion, Amazon Sponsored, etc.

performance_analysis read-only

Analyze campaign performance with actionable insights (wasted spend, anomaly detection, audience insights, creative fatigue), supports raw data mode (raw_data=true).

keyword_research read-only

Research keywords with real CPC data and competitive analysis.

budget_optimization writes

Optimize budget allocation based on performance data.

account_management writes

Manage multiple ad accounts across platforms.

scheduled_reports writes

Set up scheduled briefs, performance monitors, and cross-platform reports.

Use cases

Unified cross-platform ad campaign management

Keyword research and competitive analysis

Ad performance analysis and optimization recommendations

Automation and scheduled reporting

Supported clients

Claude Desktop

ChatGPT

Cursor

Codex

Gemini CLI

Claude Code

Windsurf

Cline

Listed from the project's documentation, not tested by this site.

Overview

A remote MCP server for creating, analyzing, and optimizing ad campaigns across Google Ads, Meta Ads, TikTok Ads, LinkedIn Ads, Amazon Ads, and ChatGPT Ads. It offers 400+ tools for campaign creation, performance analysis, keyword research, budget optimization, and more. Uses OAuth 2.1 authentication and Streamable HTTP transport. Suitable for marketers or agencies who want to manage multiple ad platforms directly from AI assistants like ChatGPT, Claude, Cursor, etc.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 62.9k · Tools 2 Compare with this →

NPM Sentinel MCP Server 79 · B

AI-powered NPM package analysis MCP server

★ 18 · Tools 19 Compare with this →

MartinLoop MCP Server 79 · B

An independent control layer that gives AI coding agents budgets, brakes, and receipts

★ 409 Compare with this →

GitHub MCP Server 76 · B

GitHub's official MCP server for managing repos, issues, PRs, and workflows via natural language

★ 33.5k · Tools 15 Compare with this →

Compare with other MCP servers

Source revision 68d3cacf6d3c Data synced 2026-10-11 Read the FMRS scoring method