ContentForge

ContentForge is an MCP server that provides AI agents and LLM clients with tools to manage the full content creation workflow, from topic research and image sourcing to SEO metadata, affiliate link enrichment, and publishing to a CMS. This server helps you streamline collaboration between components, automate repetitive tasks, and publish ready-to-use articles efficiently.

How to use

You connect to ContentForge with an MCP client using either a local stdio mode or a hosted mode that communicates over HTTP. In stdio mode you run the server as a subprocess, and in hosted mode you connect remotely to the provided endpoints. Once connected, you can orchestrate the content pipeline by activating the available functions in sequence to research, source images, optimize SEO, monetize links, and publish the finished article.

How to install

Prerequisites: you need Node.js and npm installed on your machine to run the local MCP server with stdio mode.

Step by step, run the following commands to start ContentForge in stdio mode locally, then configure your MCP client to connect.

Additional sections

Configuration and usage details you’ll use include the following server connection options and environment variables.

Available tools

list_sites

List all available sites and their categories from the CMS.

source_images

Search for stock images on Pexels with accessible URLs, alt text, and photographer credits.

seo_metadata

Generate SEO metadata including slug, meta description, schema.org markup, and internal link suggestions.

enrich_links

Scan article content for Amazon product URLs and add affiliate tags.

publish_article

Publish a complete article to the CMS with auto-categorization and tag support.

get_existing_articles

Retrieve existing articles from a site for internal linking and deduplication.