ai

This MCP server package represents the WordPress AI Experiments plugin ecosystem by provisioning a WordPress container prepped for AI experiments. The plugin itself provides modular AI features (e.g., Abilities Explorer, Alt Text Generation, Content Summarization, Image Generation, Title Generation) that developers and site admins can test and extend. With this MCP setup, you can deploy a WordPress instance and interact with the AI building blocks described in the plugin documentation. To use the server, run the provided command to start a WordPress container, then access the site at http://localhost:8080. Within WordPress, enable and configure the AI Experiment features, connect to AI providers (OpenAI, Google, Anthropic, etc.), and use the built-in experiment framework to opt-in to specific AI capabilities. The plugins’ features are exposed through Gutenberg blocks and admin screens, allowing content creators to generate alt text, summarize content, generate excerpts and titles, create images from post content, and inspect AI “abilities” via the Abilities Explorer.

Once running, you can evaluate different AI providers and configurations within WordPress, toggle features on or off, and collect feedback on performance, accuracy, and accessibility. The MCP server is meant for testing and feedback workflows, not for guaranteed production deployment; follow WordPress core and plugin guidelines when moving to production ecosystems.

Prerequisites:

• Docker installed and running on your machine or host environment
• Basic familiarity with Docker commands and WordPress setup
• Optional: a domain name and TLS if you plan to expose WordPress in production

Installation steps:

1. Ensure Docker is installed

   • macOS: install Docker Desktop
   • Windows: install Docker Desktop
   • Linux: install Docker and start the daemon

2. Pull and run the WordPress container configured for AI Experiments

   • Use the MCP server command to start the container:

   docker run -d -p 8080:80 wordpress:latest

   This will start a WordPress instance accessible at http://localhost:8080. Follow the on-screen WordPress setup wizard to configure the site (database in container may require additional steps in a real deployment).

3. Install and activate the WordPress AI Experiments plugin from within the WordPress admin

   • Log in to WordPress as an administrator
   • Navigate to Plugins > Add New and install the WordPress AI Experiments plugin, or upload the plugin if you have a local copy
   • Activate the plugin and configure AI providers, keys, and feature flags as needed

4. Configure AI builders and experiments

   • Open the AI Experiments dashboard in WordPress
   • Connect to AI providers (OpenAI, Google, Anthropic, etc.) and set API keys
   • Enable the features you want to test (Abilities Explorer, Alt Text Generation, Content Summarization, Excerpt Generation, Image Generation, Title Generation)

5. Validate and iterate

   • Create or edit content in Gutenberg and use the AI blocks to generate content, then review for accuracy and accessibility
   • Use the Review Notes feature to improve content quality

Optional advanced steps:

• Use a docker-compose workflow to orchestrate WordPress plus a database container
• Configure persistent volumes for WordPress data and database data
• Set up a reverse proxy with TLS for production deployments

Notes and tips:

• This MCP setup uses a standard WordPress docker image; for production, consider using a multi-container setup with separate MySQL/PostgreSQL containers and a reverse proxy (e.g., Nginx/Traefik).
• Ensure API keys for AI providers are stored securely (environment variables or a secrets manager) and not committed to source control.
• The AI features are experimental by design. Validate results in non-production environments and collect feedback before broader rollout.
• When exposing WordPress publicly, enforce strong authentication, update plugins/themes regularly, and monitor API usage to manage costs.
• If you need to customize the plugin behavior, consult the WordPress AI Experiments development docs and the Abilities API documentation.
• If running locally, remember that port 8080 maps to the container’s port 80; adjust firewall rules accordingly.