meta-ads-mcp

MCP安全风控

提供方 andresochoa1

安全风控 stdio · 本地进程 API Key 社区 / 第三方 Claude DesktopClaudeCursor
查看接入方式

A security-first MCP server for the Meta Ads API that provides 18 tools for accounts, campaigns, ad sets, ads, creatives, and insights, enabling secure management and analysis of Facebook and Instagram ads via MCP.

一句话选型结论

按资料字段整理

meta-ads-mcp 属于安全风控类 MCP 服务器,stdio · 本地进程,API Key,可直接接入 Claude Desktop、Claude、Cursor,有可直接执行的安装命令。接入前先确认权限范围;涉及写入、删除或付款的操作,建议先在测试环境验证。

分类安全风控 传输方式stdio · 本地进程 鉴权API Key 部署方式本地部署
GitHub 星标 0 仓库热度,无公开仓库时显示未标注
工具数 未标注 暴露给模型的工具数量
资源 / 提示词 0 / 0 资源与提示词模板数量
维护方 社区 / 第三方 核验等级 仅资料核对

工具与能力

安全风控
数量来自来源目录的标注,具体工具清单以官方文档为准
工具数量未标注暴露给模型可调用的工具个数
资源数量0可读取的资源条目
提示词数量0内置提示词模板数量

能力标签

Local Deployment Ad Management Meta Graph API Python MCP Protocol Zero Trust Security Ads Performance Analysis Ad Account Management Facebook Ads Instagram Ads

支持客户端

适用场景

Local DeploymentAd ManagementMeta Graph API

安装与配置

stdio · 本地进程 API Key
客户端在本机拉起子进程,数据不出本机
meta-ads-mcp

这条命令只是入口,具体启动参数请按官方文档补齐,不要直接照搬到别的服务器。

  1. 1
    准备运行环境

    stdio 服务在本地起进程,需要先装好 python,再按安装命令拉取依赖。

  2. 2
    安装服务

    复制上面的安装命令执行;命令里带占位符的位置,换成自己的密钥或路径。

  3. 3
    写入客户端配置

    在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,让新配置生效。

  4. 4
    验证是否可用

    连接成功后随便调一个只读工具,确认能返回结果,再接入正式流程。

典型使用场景

3 类
按分类与标签整理的常见用法,具体以官方文档为准
资产与漏洞排查

盘点资产、扫漏洞、看暴露面。

日志与审计

汇总日志、查异常行为。

合规检查

按规则检查配置与权限是否合规。

选择建议

选之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 有现成的安装命令,不用自己找入口,照着跑就能起服务。
  • 已适配 Claude Desktop、Claude、Cursor,换客户端不用重新研究配置。

先想清楚的地方

  • stdio 方式要求本地装好运行时(如 Node.js),换机器要重新装一遍。
  • 使用前要准备好API Key对应的凭据,并按最小权限配置。
  • 资料没标注工具数量,实际能力要装好后确认。
  • 版本更新与行为细节以官方仓库为准。

更适合

  • 需要在安全风控场景里让模型直接调用外部能力的人。
  • 已经在用 Claude Desktop、Claude 的用户,接入路径最短。
  • 想先快速试一下、能接受在本地或自有环境跑服务的人。

不太适合

  • 完全不想在本地装环境的人。
  • 不方便管理密钥或第三方授权的人。
  • 需要厂商级 SLA 和技术支持的生产环境。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
MCP 服务器工具数传输方式部署方式安装命令最后核验
meta-ads-mcp当前 未标注 stdio · 本地进程 本地部署 有 2026-10-29
Cyber Strike AI 75 未标注 未标注 无 2026-05-17
mcp-for-beginners 未标注 stdio · 本地进程 本地部署 有 2026-04-13
casdoor 未标注 stdio · 本地进程 本地部署 有 2026-04-12
kubeshark 未标注 stdio · 本地进程 本地部署 有 2026-04-11
Modelcontextprotocol C22988b0 未标注 stdio · 本地进程 本地部署 有 2026-05-17

服务器档案

安全风控
字段来自项目仓库与公开索引,中文说明由本站标注
提供方andresochoa1上游标注的维护方 andresochoa1
分类安全风控本站按上游分类展示
连接方式stdio · 本地进程客户端在本机拉起子进程,数据不出本机
鉴权方式API Key远程服务建议用最小权限凭据
维护方社区 / 第三方核验等级 仅资料核对
包名 / 运行时未标注 · python本地安装时用到的包标识与运行时
部署类型本地部署上游给出的部署形态
最近核验2026-10-29上游最后核验时间

详细介绍

9.5 KB
来自项目仓库与上游目录的原文,未做删改

meta-ads-mcp

Security-first MCP server for Meta (Facebook/Instagram) Ads API.


Why this exists

Existing MCP integrations for Meta Ads have significant security gaps: no SSRF protection on pagination URLs, access tokens leaking into logs and error messages, no rate limiting, and no input validation. This server is built from scratch with zero-trust principles. Every outbound request is validated against a domain allowlist, every log line is automatically redacted, and every ID is pattern-checked before it touches the API.

Features

  • 18 tools covering accounts, campaigns, ad sets, ads, creatives, and insights
  • 3 MCP resources for campaign metadata and account summaries
  • 5 built-in MCP prompts for common analysis workflows (performance analysis, underperformer detection, budget efficiency, creative fatigue, daily monitoring)
  • Zero-trust security -- SSRF protection via domain allowlist, token redaction in all logs and errors, input validation on all IDs
  • Rate limiting with token bucket algorithm and exponential backoff with jitter
  • Structured logging with automatic secret redaction (structlog)
  • Built on Meta's Graph API via httpx with zero-trust HTTP controls

Quick Start

Prerequisites

  • Python 3.11 or higher
  • A Meta Developer account with an app that has ads_read permission
  • A valid Meta Graph API access token

Install```bash

pip install meta-ads-mcp


### Get your access token

1. Go to Meta Developer Tools
2. Select your app
3. Generate a token with `ads_read` permission
4. For production use, exchange for a long-lived token

### Set environment variable```bash
export META_ACCESS_TOKEN=your_token_here

Configure in Claude Desktop

Add to your Claude Desktop claude_desktop_config.json:```json
{
"mcpServers": {
"meta-ads": {
"command": "meta-ads-mcp",
"env": {
"META_ACCESS_TOKEN": "your_token_here"
}
}
}
}


### Configure in Claude Code

Add to your project or global `.claude/settings.json`:```json
{
  "mcpServers": {
    "meta-ads": {
      "command": "meta-ads-mcp",
      "env": {
        "META_ACCESS_TOKEN": "your_token_here"
      }
    }
  }
}

Configuration

All configuration is via environment variables. The access token is never accepted as a CLI argument.

Variable Required Default Description
META_ACCESS_TOKEN Yes -- Meta Graph API access token
META_API_VERSION No v22.0 Graph API version
META_RATE_LIMIT_PER_HOUR No 200 Maximum API calls per hour
LOG_LEVEL No INFO Logging level (DEBUG, INFO, WARNING, ERROR)
DEBUG No false Enable debug mode with console rendering

Available Tools

Accounts (3)

Tool Description Required Parameters
list_ad_accounts List all accessible ad accounts with name, ID, status, currency, balance, and spend --
get_ad_account_details Get detailed info for a specific ad account account_id
get_ad_account_activities Get the activity log for an ad account account_id

Campaigns (3)

Tool Description Required Parameters
list_campaigns List campaigns for an account, filterable by status account_id
get_campaign Get campaign details including objective, budget, and schedule campaign_id
get_campaign_insights Get performance metrics with date presets, breakdowns, and time increments campaign_id

Ad Sets (4)

Tool Description Required Parameters
list_adsets_by_account List ad sets for an account, filterable by status account_id
list_adsets_by_campaign List ad sets for a specific campaign campaign_id
get_adset Get ad set details including targeting, budget, and optimization goal adset_id
get_adset_insights Get performance metrics for a specific ad set adset_id

Ads (4)

Tool Description Required Parameters
list_ads_by_account List ads for an account, filterable by status account_id
list_ads_by_campaign List ads for a specific campaign campaign_id
list_ads_by_adset List ads for a specific ad set adset_id
get_ad_insights Get performance metrics for a specific ad ad_id

Creatives (2)

Tool Description Required Parameters
get_creative Get creative details including title, body, image URL, and object story spec creative_id
list_creatives_by_ad List creatives associated with a specific ad ad_id

Insights (2)

Tool Description Required Parameters
get_account_insights Get comprehensive account-level insights with all breakdown and attribution options account_id
fetch_next_page Fetch the next page of results from a pagination URL (SSRF-validated) url

Common Optional Parameters

Most tools support these optional parameters:

Parameter Type Description
fields string[] Specific fields to retrieve
effective_status string[] Filter by status (ACTIVE, PAUSED, ARCHIVED, etc.)
limit integer Maximum results per page (default: 25)
date_preset string Date range preset (today, yesterday, last_7d, last_30d, this_month, last_month)
time_range object Custom date range with since and until (YYYY-MM-DD)
time_increment string Time granularity (1, 7, 28, monthly, all_days)
breakdowns string[] Breakdown dimensions (age, gender, country, placement, etc.)

Available Resources

URI Name Description
meta://accounts Ad Accounts List of all accessible ad accounts with basic info
meta://accounts/{account_id}/campaigns Account Campaigns Active campaigns for a specific account
meta://accounts/{account_id}/summary Account Summary Executive summary with account details and last 7 days performance

Available Prompts

Prompt Description Required Arguments
analyze-campaign-performance Analyze metrics and provide optimization recommendations campaign_id
find-underperformers Identify ads performing below CPA/ROAS targets account_id
daily-monitoring-report Generate a daily health report with spend pacing and anomaly detection account_id
budget-efficiency-check Evaluate budget distribution vs. ROAS performance account_id
creative-fatigue-audit Detect ads with declining CTR, rising frequency, and stale engagement account_id

Security Model

This server implements a zero-trust security architecture. Key protections include:

  • SSRF protection -- All outbound URLs (including pagination) are validated against a strict domain allowlist (graph.facebook.com, www.facebook.com)
  • Token management -- Access tokens are loaded from environment variables only, never accepted as CLI arguments, and never appear in logs or error messages
  • Input validation -- All Meta Ads IDs are validated against expected patterns before API calls
  • Log redaction -- A structlog processor automatically redacts sensitive fields (access_token, secret, password, key, authorization) from all log output
  • No redirects -- HTTP client is configured to never follow redirects, preventing redirect-based SSRF
  • HTTPS only -- All requests are validated to use HTTPS

For the complete security documentation, see SECURITY.md.

Development

Clone and install```bash

git clone https://github.com/andresochoa1/meta-ads-mcp.git
cd meta-ads-mcp
pip install -e ".[dev]"


### Run tests```bash
pytest

Lint```bash

ruff check src/ tests/
ruff format --check src/ tests/


### Project structure```
src/meta_ads_mcp/
  server.py          # MCP server entry point, tool/resource/prompt registration
  config.py          # Environment-based configuration with validation
  security.py        # SSRF protection, token redaction, input validation
  api/
    client.py        # Secure Meta Graph API client
    rate_limiter.py  # Token bucket + exponential backoff
    pagination.py    # Cursor-based pagination with URL validation
  tools/
    accounts.py      # Ad account tools
    campaigns.py     # Campaign tools
    adsets.py        # Ad set tools
    ads.py           # Ad tools
    creatives.py     # Creative tools
    insights.py      # Insights and pagination tools
  prompts/
    analysis.py      # Built-in analysis prompts
  resources/
    campaign_data.py # MCP resource implementations
tests/
  test_security.py   # Security layer tests
  test_rate_limiter.py # Rate limiter tests
  test_pagination.py # Pagination tests

License

MIT -- Copyright (c) 2026 Andres Ochoa

Contributing

Contributions are welcome. Please follow these guidelines:

  1. Fork the repository and create a feature branch
  2. Write tests for any new functionality
  3. Run the full test suite (pytest) and linter (ruff check) before submitting
  4. Follow the existing code style -- the project uses ruff for formatting and linting
  5. Security-sensitive changes require extra review. If your change touches security.py, client.py, or URL handling, please explain the security implications in your PR description
  6. Open an issue first for large changes to discuss the approach before implementation

Reporting security vulnerabilities

Please do not open a public issue for security vulnerabilities. Instead, see SECURITY.md for responsible disclosure instructions.

常见问题

8 问
答案由本站依据条目字段整理,不引用上游原文
meta-ads-mcp 怎么安装?

在客户端按下面的安装命令接入:meta-ads-mcp。装完在客户端配置里引用同一个包或端点。

meta-ads-mcp 怎么在 Claude Desktop、Claude、Cursor 里配置?

它的传输方式是stdio · 本地进程,在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,配置才会生效。

meta-ads-mcp 收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于服务本身和背后的接口额度,正式使用前按官方说明确认。

meta-ads-mcp 需要 API Key 吗?

资料标注需要API Key认证,使用前准备好对应凭据,并按最小权限配置。

meta-ads-mcp 支持哪些客户端?

资料标注支持 Claude Desktop、Claude、Cursor;其他客户端只要能配置 MCP Server,也能用同样的命令或地址接入。

meta-ads-mcp 能提供哪些能力?

资料显示能力标签包括 Local Deployment、Ad Management、Meta Graph API、Python、MCP Protocol、Zero Trust Security。具体工具清单以官方文档为准。

meta-ads-mcp 需要什么运行环境?

运行时为 python;属于本地进程,需要在本机装好运行环境。

meta-ads-mcp 是本地还是远程?

它是本地进程(stdio),由客户端在本机拉起。传输方式为stdio · 本地进程。

热门专题

按标签挑服务器
全部专题 →

服务器信息以作者仓库与官方文档为准;分类、传输方式与鉴权方式的中文说明为中文整理,最近核验 2026-10-29。