Marketic — Marketing Intelligence OS

AI-native marketing operating system that encodes real performance marketing judgment.

Closed-loop: Analyze competitors → Generate counter-variants → Launch campaigns → Measure → Learn

---

🎯 What Marketic Does

# Start the MCP server
python3 mcp_server.py

# In another terminal, analyze a competitor
echo '{"jsonrpc":"2.0","method":"tools/call","params":{"name":"analyze_competitor","arguments":{"brand":"HubSpot","category":"marketing automation"}},"id":1}' | python3 mcp_server.py

Output:

{
  "competitor": "HubSpot",
  "analysis": "HubSpot dominates with all-in-one platform...",
  "confidence": 0.85,
  "analyzed_at": "2026-07-20T14:15:40.171150"
}

Why this is different: Marketic exposes 32 marketing intelligence tools via MCP (Model Context Protocol), enabling any MCP-compatible agent (including Quay) to call marketing domain functions natively.

---

🔌 MCP Server — 32 Tools

Marketic exposes 32 tools across 7 categories via JSON-RPC 2.0 over stdio:

Category	Tools	Use Case
Competitive Analysis	analyze_competitor, compare_competitors, analyze_positioning	Competitor research, gap analysis, positioning maps
Creative Generation	generate_creatives, generate_social_posts, generate_seo_content, generate_narrative	Ad copy, social posts, SEO articles, brand stories
Campaign Management	build_campaign, optimize_budget, launch_campaign_ad	Multi-channel campaign planning, ROAS optimization
Analytics	get_attribution, collect_signals	Attribution modeling, market signal collection
Hub Connectors	hub_* (9 tools)	Unified marketing platform integration (WebEngage, HubSpot, etc.)
CRM	crm_* (8 tools)	Lead/deal management, activity logging
Utilities	build_utm_url, parse_utm_params, run_workflow	URL tracking, workflow orchestration

Quick Test

# List all tools
echo '{"jsonrpc":"2.0","method":"tools/list","params":{},"id":1}' | python3 mcp_server.py

# Run a sample campaign suite
python3 run_sample_campaigns.py

---

🏗️ Architecture

Signal → Plan → Execute → Respond → Learn
   ↓        ↓        ↓         ↓        ↓
 Context  Route   Ensemble   Merge    Quality
 Loading   Tier   +Fallback  +Deliver  Tracking

Tiered LLM Routing:

• Easy task (ad copy): 1 cheap model, ~$0.0001
• Medium task (gap analysis): 2 mid-tier models parallel, ~$0.002
• Hard task (campaign strategy): 3 premium models + confidence voting, ~$0.05

---

🚀 Integration with Quay

Marketic is integrated into Quay (Autonomous AI Software Factory) as a native MCP tool provider:

// Quay's marketing module config (src/server/marketing/config.ts)
export const MARKETIC_MCP_CONFIG: MCPServerConfig = {
  name: 'marketic',
  command: 'python3',
  args: ['/Users/Subho/marketic/mcp_server.py'],
  env: { /* API keys */ },
};

Quay Marketing Routes:

• POST /api/marketing/analyze → marketic::analyze_competitor
• POST /api/marketing/creatives → marketic::generate_creatives
• POST /api/marketing/campaign → marketic::build_campaign / launch_campaign_ad
• POST /api/marketing/signals → marketic::collect_signals

---

📊 Test Results (2026-07-20)

67 integration tests passing across Marketic MCP server and Quay API routes:

npm test -- tests/integration/
# 43 Marketic MCP tool tests + 24 API route contract tests

Sample Campaigns Verified

Campaign	Tools Chained	Status
A: Position-and-Attack	analyze_competitor → analyze_positioning → generate_creatives	✅
B: ROAS Optimizer	get_attribution → optimize_budget	✅
C: Creative-Bomb	generate_creatives → generate_social_posts → build_campaign	✅
D: SEO Content Engine	generate_seo_content → generate_narrative	✅

---

🛠️ Quick Start

# Clone
git clone https://github.com/Das-rebel/marketic
cd marketic

# Install dependencies
pip install -e .

# Run MCP server (for use by Quay or other MCP clients)
python3 mcp_server.py

# Or run sample campaigns
python3 run_sample_campaigns.py

---

📁 Project Structure

marketic/
├── mcp_server.py           # MCP stdio server (32 tools)
├── run_sample_campaigns.py  # Sample campaign runner
├── API_DOCUMENTATION.md    # Full API reference
├── marketic/
│   ├── gtm/                # Go-to-market tools
│   │   ├── positioning.py
│   │   ├── narrative.py
│   │   └── ...
│   ├── creative/           # Creative generation
│   │   ├── social_generator.py
│   │   ├── seo_generator.py
│   │   └── ...
│   ├── campaign/           # Campaign management
│   │   ├── builder.py
│   │   ├── budget_router.py
│   │   └── ...
│   ├── analytics/          # Attribution & signals
│   ├── integrations/       # Hub connectors
│   └── crm/               # CRM tools
└── tests/                 # Integration tests

---

⚙️ Environment Variables

# LLM API keys (for real AI-powered analysis)
OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...

# Marketing platform integrations (optional)
HUBSPOT_API_KEY=...
WEBENGAGE_API_KEY=...
CLEVERTAP_ACCOUNT_ID=...
BRAZE_API_KEY=...
MAILCHIMP_API_KEY=...
CLAY_API_KEY=...

---

📈 Differentiation

Tool	What it does	Marketic advantage
Icon AI CMO	Black-box autonomous ads	Transparency: every decision logged with reasoning
HubSpot/Marketo	Enterprise platforms	Open, learnable system that gets smarter over time
dub.co	Link attribution only	Attribution + competitive intel + creative generation
claude-seo	SEO skills only	Full marketing intelligence, not just SEO

---

📄 Documentation

• API_DOCUMENTATION.md — Complete 32-tool API reference with examples
• AGENT_COUNCIL_REPORT.md — Architectural review by multiple AI agents

---

License

MIT © Subhajit Das