awesome-agentic-advertising

Awesome Agentic Advertising is a curated catalog of MCP servers and resources for automating advertising campaigns with AI agents. It aggregates cross-platform tools and examples for Google Ads, Meta Ads, LinkedIn Ads, Amazon, TikTok, and more, along with related protocols and agent frameworks. Use this directory as a reference to discover MCP servers you can deploy and integrate into an agent-driven workflow for planning, creating, optimizing, and reporting on ad campaigns.

To use the contents, browse the listed sections to identify relevant MCP servers (e.g., cross-platform tools or platform-specific servers) and follow their individual documentation for installation and integration. Many entries link to GitHub repositories or official sites where you can obtain the server, review capabilities (campaign creation, AI-generated creatives, budget optimization, data connectivity, conversions tracking), and learn how to connect them to your agent framework or automation pipeline. When integrating multiple MCP servers, leverage MCP-compatible prompts and AdCP/A2A standards to enable coordinated agent actions across platforms.

If you’re building an autonomous advertising workflow, start by selecting a cross-platform MCP server to handle orchestration and data normalization, then pair it with platform-specific MCPs for execution, analytics, and conversion tracking. Ensure you understand authentication requirements, API limits, and data privacy considerations, and keep an eye on the respective project’s licensing and contribution guidelines as you experiment and extend your toolchain.

Prerequisites:

• A supported runtime environment (Node.js, Python, or Docker, depending on the MCP servers you intend to use).
• Git installed to clone repositories and access example configurations.
• Access to any required API credentials for advertising platforms (e.g., Google Ads API, Meta Marketing API).

Step-by-step:

1. Clone this repository to your local machine or a workspace: git clone https://github.com/jshorwitz/awesome-agentic-advertising.git cd awesome-agentic-advertising

2. Review the contents to identify MCP servers you want to deploy. Each server typically has its own README with installation instructions. For example, if you choose a cross-platform MCP server, follow its specific setup guide in its repository.

3. Install prerequisites for the selected MCP server(s). Common steps may include:

   • npx-based servers: ensure Node.js is installed and run the server via its npm package (e.g., npx -y <package-name>).
   • Python-based servers: set up a virtual environment and install requirements.txt.
   • Docker-based servers: install Docker and pull the appropriate image.

4. Configure environment variables and API credentials as required by the chosen MCP server(s). Common variables include client IDs, secrets, and base API URLs. Create a .env file or export variables in your environment as instructed by the server docs.

5. Run the MCP server(s) according to their instructions. Verify connectivity and test basic prompts to confirm the server responds as expected.

6. Integrate into your agent workflow by establishing prompts, adapters, and data flows between MCP servers, ensuring you follow the MCP and AdCP standards used by the ecosystem.

Notes and tips:

• This repository is a curated list of MCP servers and resources, not a single runnable server. Each entry has its own setup and usage instructions.
• When integrating multiple MCP servers, consider using a cross-platform orchestrator to coordinate tasks and normalize data across platforms.
• Pay attention to API rate limits, authentication scopes, and permissions for each advertising platform you connect to.
• Some entries may be in active development or have beta features; review their commit history and licensing before production use.
• Keep credentials secure and rotate API keys regularly. Use environment variables or secret management solutions for deployments.
• Check for updates in MCP standards (MCP, AdCP, A2A) to ensure compatibility when wiring multiple servers together.