AI Marketing Growth OS - Agentic Marketing Platform

An agentic marketing operating system with governed AI agents.

No hosted demo for this project — it runs locally. See install instructions below.

AI & Automation

Overview

AI Marketing Growth OS is a Python/FastAPI agentic platform that orchestrates specialized AI agents to plan, execute and monitor marketing operations. It ships a FastAPI backend, a single-page dashboard, Docker Compose setup and an evaluation lab, with human-in-the-loop approval for any write-capable action.

Problem It Solves

Marketing teams increasingly rely on AI outputs that are ungoverned, untraceable and disconnected from campaign operations. Growth decisions need agentic automation with least-privilege tool access, audit trails and mandatory human approval before any budget-affecting action is executed.

Target Users

Marketing leaders and growth teams
Agencies running multi-channel campaigns
MarTech engineers evaluating agent orchestration

Key Features

Four specialized agents with bounded, least-privilege tool registries
Human-in-the-loop approval state machine for write-capable actions
Model Context Protocol (MCP) adapter for tool interop
Evaluation lab with benchmark ground truths
Docker Compose for containerized deployment
Zero-cost standalone demo (python demo.py) with no API keys required

Technologies

Python
FastAPI
Docker
Model Context Protocol (MCP)
Pydantic
SQLite
Chart.js / Tailwind dashboard

Contribution & Skills

Sole author. Designed the agent architecture with permission-scoped tool registries, built the orchestrator with a governance state machine, implemented prompt-injection sanitization for external inputs and shipped the evaluation lab and containerized deployment.

Install & Usage

Quick demo

git clone https://github.com/hamid-digitalmarketer/ai-marketing-growth-os.git
cd ai-marketing-growth-os
python demo.py

Local development

pip install -r requirements.txt
uvicorn backend.app.main:app --reload --port 8000

Containerized

docker compose up --build

Full documentation lives in the repository README →

FAQ

Is AI Marketing Growth OS a hosted service?

No. The project runs locally or in your own Docker environment. There is no hosted public demo; the repository includes a standalone demo script that runs without external API keys.

How does human-in-the-loop governance work?

Write-capable tools are isolated and intercepted by the orchestrator. Any action must transition through PENDING, APPROVED, EXECUTED and MEASURED states, and execution is halted while an action is pending approval.