fb-ads-automation
Enables full read/write management of Facebook ad campaigns, ad sets, ads, and creatives via the Meta Marketing API through natural language, with AI creative generation, performance analytics, and PDF reporting.
README
Facebook Ads Automation
An MCP server for Claude Code that gives you full read/write access to the Meta Marketing API. Manage campaigns, analyze performance, generate AI creatives, and get optimization recommendations — all through natural language.
77 MCP tools | 7 AI agents | PDF reports | AI creative generation | Multi-account support
Built on FastMCP + Facebook Marketing API v24.0.
How It Works
YOU (natural language in Claude Code)
|
v
SLASH COMMANDS /fb-ads:daily-check, /fb-ads:weekly-report, /fb-ads:optimize, ...
|
v
AGENTS (7) Report, Market Analyst, Creative Strategist, Creative Generator,
Performance Optimizer, Competitive Intel, Ads Optimizer
|
v
KNOWLEDGE Per-client context: business goals, brand voice, audience, competitors
|
v
MCP SERVER (77) Facebook API tools + Analytics engine + AI creative generation
What You Can Do
- Manage campaigns — Create, read, update, pause/activate campaigns, ad sets, ads, and creatives
- Analyze performance — Get insights, run analytics, detect creative fatigue, check audience saturation
- Generate creatives — AI image/video generation via Freepik (14 image + 24 video models), Gemini (Imagen 3 + Veo), and OpenAI (DALL-E 3 + Sora)
- Edit images — Upscale, remove backgrounds, reimagine, recolor, inpaint, and more (9 editing operations)
- Optimize — Get data-driven recommendations based on your KPIs, thresholds, and business rules
- Generate reports — KPMG-style PDF reports with charts, KPIs, and recommendations
- Research competitors — Search the Facebook Ad Library and analyze competitor strategies
- Manage multiple accounts — Switch between ad accounts with labeled aliases
Quick Start
1. Install
git clone https://github.com/your-username/fb-ads-automation.git
cd fb-ads-automation
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
2. Configure
cp .env.example .env
cp .mcp.json.example .mcp.json
Edit .env with your Facebook credentials (see Setup Guide for step-by-step instructions on getting each credential):
FB_APP_ID=your_app_id
FB_APP_SECRET=your_app_secret
FB_ACCESS_TOKEN=your_access_token
FB_AD_ACCOUNT_ID=act_your_account_id
Edit .mcp.json with absolute paths to your Python and server:
{
"mcpServers": {
"fb-ads": {
"command": "/absolute/path/to/.venv/bin/python3",
"args": ["/absolute/path/to/fb_ads_mcp_server/server.py"]
}
}
}
3. Verify
Open Claude Code in the project directory:
check_token_status
If you see your token info, you're ready.
4. Onboard Your First Client
/fb-ads:onboard-client
This interactive wizard validates your account, interviews you about your business, and creates the knowledge files that agents use.
Tool Groups
Only 8 core tools load at startup. Load more on demand:
enable_tools(groups="crud,insights") # Campaign management + reporting
enable_tools(groups="creative_gen") # AI image/video generation
enable_tools(groups="all") # Everything (77 tools)
| Group | Tools | What It Does |
|---|---|---|
| core | 8 | Account management, token tools, tool loading |
| crud | 21 | Create/read/update campaigns, ad sets, ads, creatives |
| insights | 5 | Performance data and breakdowns |
| analytics | 7 | Trends, fatigue detection, saturation analysis |
| targeting | 3 | Interest search, audience estimation, suggestions |
| pages | 5 | Facebook Page management and insights |
| 4 | Instagram account and media management | |
| ad_library | 2 | Competitor ad research |
| creative_gen | 7 | AI image and video generation |
| image_editing | 9 | Upscale, remove BG, reimagine, recolor, inpaint |
| stock | 3 | Search Freepik image/icon/video libraries |
| audio | 3 | AI music, sound effects, vocal isolation |
Full tool reference: docs/mcp-tools.md
Agents
Seven specialized agents live in agents/. Activate by saying "Load [agent name]":
| Agent | What It Does |
|---|---|
| Report Agent | Generates KPMG-style PDF reports with charts and KPIs |
| Market Analyst | Audience research, demographic analysis, expansion planning |
| Creative Strategist | Ad copy, messaging angles, A/B test design |
| Creative Generator | AI image/video generation with brand-aware prompts |
| Performance Optimizer | Data-driven optimization recommendations |
| Competitive Intel | Ad Library research, competitor strategy analysis |
| Ads Optimizer | Campaign-level optimization (deprecated, kept for reference) |
Agents auto-read your knowledge files for business context, brand voice, audience data, and competitive landscape.
Slash Commands
| Command | What It Does |
|---|---|
/fb-ads:onboard-client |
Set up a new client (interactive wizard) |
/fb-ads:daily-check |
Morning health check on your ad account |
/fb-ads:weekly-report |
Weekly performance summary |
/fb-ads:optimize |
Analyze and get optimization recommendations |
/fb-ads:new-campaign |
Interactive campaign creation wizard |
/fb-ads:creative-refresh |
Check for creative fatigue + guide refresh |
/fb-ads:generate-creative |
Generate AI images/videos for ads |
/fb-ads:competitor-check |
Research competitor ads via Ad Library |
Knowledge System
Agents use per-client knowledge files to understand your business:
knowledge/
playbook.md # Global optimization thresholds
accounts.md # Account registry
report-specs.md # PDF report configuration
clients/
_templates/ # Blank templates to copy
_example/ # Filled example (fictional cafe)
your-client/ # Your client's knowledge files
business-context.md # Company, product, KPIs, strategy
brand-voice.md # Tone, messaging, visual direction
audience-research.md # Segments, demographics, saturation
competitive-landscape.md # Competitors, benchmarks, gaps
Important Notes
Budgets use the smallest currency unit. USD uses cents ($50/day = daily_budget: 5000). INR uses paisa (Rs 1,000/day = daily_budget: 100000).
Campaign objectives must be OUTCOME-based: OUTCOME_TRAFFIC, OUTCOME_ENGAGEMENT, OUTCOME_LEADS, OUTCOME_AWARENESS, OUTCOME_SALES, OUTCOME_APP_PROMOTION.
Rate limits are ~200 API calls/hour/user. Analytics tools make multiple calls per invocation.
Tokens expire after ~60 days. Use check_token_status and refresh_access_token to manage them.
Optional: AI Creative Generation
The core ads management works without these. Add API keys for creative generation:
| Provider | Models | Key Required |
|---|---|---|
| Freepik | 14 image + 24 video + editing | FREEPIK_API_KEY |
| Google Gemini | Imagen 3 + Veo | GEMINI_API_KEY |
| OpenAI | DALL-E 3 + Sora | OPENAI_API_KEY |
See docs/freepik.md for model selection guides.
Documentation
| Doc | Contents |
|---|---|
| Setup Guide | Step-by-step installation and credential setup |
| MCP Tools Reference | All 77 tools by group with descriptions |
| Agents & Clients | Agent system, onboarding, code structure |
| Freepik Guide | AI models, editing operations, parameters |
| Dashboard | Client dashboard architecture and deployment |
Project Structure
fb_ads_mcp_server/ # MCP server
server.py # Entry point (FastMCP)
api/ # Pure Facebook SDK wrappers (no MCP coupling)
tools/ # MCP tool registration (thin orchestrators)
analytics/ # Pure computation functions
rules/ # Business rule engine
providers/ # AI provider abstraction (Gemini, OpenAI, Freepik)
agents/ # 7 agent instruction files
knowledge/ # Global + per-client knowledge files
reports/ # PDF report generator + charts
dashboard/ # Static client dashboard (Netlify-ready)
docs/ # Documentation
tests/ # Test suite
Contributing
See CONTRIBUTING.md for development setup, testing, and PR guidelines.
License
MIT - see LICENSE.
Recommended Servers
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.
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.
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.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
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.
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.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
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.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.