Marketic MCP Server

32 marketing intelligence tools exposed via the Model Context Protocol (MCP) for AI coding assistants.

What is this?

A standalone MCP server that gives AI coding assistants real marketing superpowers:

• Competitor Analysis - Analyze any brand's positioning, pricing, marketing approach
• Ad Creative Generation - Generate 5 variants of ad copy for any channel (Meta, Google, LinkedIn, etc.)
• Campaign Planning - Build multi-channel campaigns with budget allocation
• Social Media - Generate platform-specific posts for LinkedIn, Twitter, Instagram
• SEO Content - Create SEO-optimized blog posts and landing page copy
• Attribution Modeling - Multi-touch attribution with Linear, Time-Decay, Position-Decay
• Budget Optimization - Route budget to best-performing channels using ROAS data
• HubSpot CRM - Full CRM: contacts, deals, workflows, email campaigns
• Signal Collection - Aggregate trends from Product Hunt, Twitter, Google Trends

Installation

npm install -g marketic-mcp

Quick Start

# Start the MCP server (stdio mode)
marketic-mcp

# Test it directly
echo '{"jsonrpc":"2.0","method":"tools/call","params":{"name":"analyze_competitor","arguments":{"brand":"HubSpot","category":"marketing automation"}},"id":1}' | python3 mcp_server.py

MCP Tools (32 total)

| Category | Tools | |----------|-------| | Competitor Intel | analyze_competitor, compare_competitors | | Creative Generation | generate_creatives, generate_social_posts, generate_seo_content | | Campaign Management | build_campaign, optimize_budget, run_pacing_model | | Attribution | run_attribution_model, calculate_roi, simulate_funnels | | Analytics | analyze_performance, ab_analyze, segment_audience, predict_ltv | | HubSpot CRM | hub_sync_contact, hub_broadcast_event, hub_send_campaign, hub_search_prospects, hub_create_segment, hub_send_transactional | | General CRM | crm_create_lead, crm_create_deal, crm_move_deal, crm_log_activity | | Workflows | run_workflow | | Collection | collect_signals | | Optimization | optimize_ctr, model_churn_risk, credit_allocate |

Configuration

# Optional: a3m-router for LLM calls (faster, cached)
export A3M_BASE_URL="http://localhost:8787/v1"
export GROQ_API_KEY="gsk_..."

# Or direct Groq
export GROQ_API_KEY="gsk_..."

Use with Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "marketic": {
      "command": "marketic-mcp"
    }
  }
}

Use with Cursor

Settings → MCP Servers → Add new:

Name: Marketic
Command: marketic-mcp

Use with Windsurf

Windsurf Settings → MCP → Add server:

Name: Marketic  
Command: marketic-mcp

Architecture

AI Assistant (Claude/Cursor/Windsurf)
        ↓ MCP stdio
marketic-mcp (npm package)
        ↓
Python MCP Server → LLM Router → Groq/a3m-router
                         ↓
               Marketic Foundation Models
               (Attribution, Budget, Analytics, Copy)

License

MIT - github.com/Das-rebel/marketic