MCP Servers

A collection of Model Context Protocol servers, templates, tools and more.

AI-Marketing-OS-MCP

Created 4/4/2026
Updated about 6 hours ago
Repository documentation and setup instructions

AI Marketing OS (MCP)

Codename: AI-Marketing-OS-MCP

Marketing MCP Server is a remote Model Context Protocol (MCP) server for content automation: research from RSS, multi-platform drafts, publishing hooks, analytics, and Notion reporting. It is meant to work with Claude and other MCP clients over HTTP + SSE.

| Property | Value | |----------|-------| | npm package | marketing-mcp-server | | MCP server name | marketing-content-automation | | Node.js | 20+ |

Claude project (the other half)

This repo is Layer 3 — the MCP server that handles research storage, drafts, publishing, and Notion analytics over HTTP + SSE.

The Layer 1 setup lives in Claude: a Project with your brand context files, platform prompts, examples, and workflow. To build that side of AI Marketing OS, follow the full guide:

AI Marketing OS — build your own automated content system (thebuilder.company)

Configure the Claude project first, then connect this server as a remote MCP (see below).

What it does

The server exposes MCP tools that drive a workflow backed by MongoDB:

| Tool | Purpose | |------|---------| | execute_research | Pulls headlines from curated RSS feeds (Reuters, Bloomberg, FT), stores a ResearchLog, returns a summary. | | save_draft_content | Saves newsletter, LinkedIn variants, X thread, and optional Instagram copy as a ContentDraft (status draft). | | publish_approved_content | Loads a draft by ID and publishes to X, LinkedIn, and Beehiiv (implementation is scaffolded; enable when API keys are set). | | sync_analytics_to_notion | Aggregates metrics (mock placeholders until real platform APIs are wired), stores AnalyticsEntry documents, and creates a report page in Notion via the official SDK. |

Also included: an Express HTTP API, SSE transport for MCP, optional OAuth discovery endpoints for remote MCP clients, and optional Bearer auth via MCP_AUTH_SECRET.

Quick start

git clone https://github.com/YOUR_ORG/YOUR_REPO.git
cd YOUR_REPO
npm install
cp .env.example .env
# Edit .env: set at least MONGODB_URI; add NOTION_* and MCP_AUTH_SECRET as needed
npm run build
npm start

Run TypeScript directly during development:

npm run dev

By default the server listens on PORT or 3000. Health check: GET /health.

Configuration

Copy .env.example to .env and fill in values. Do not commit .env (it is listed in .gitignore).

  • MongoDB - MONGODB_URI (local or Atlas).
  • MCP - optional MCP_AUTH_SECRET. When set, clients must send Authorization: Bearer <secret> (or x-mcp-auth) for protected routes such as /sse and /messages.
  • Notion - NOTION_API_KEY, NOTION_DATABASE_ID for analytics sync. Your Notion database properties should match the names used in src/index.ts, or adjust the code.
  • Publishing - X/Twitter, LinkedIn, and Beehiiv variables as described in .env.example.

For production hosting (e.g. EC2, HTTPS, process manager), see DEPLOYMENT.md.

Remote MCP (Claude)

After your Claude Project files are in place, add this server under Settings → Integrations using your deployed HTTPS base URL. The app exposes OAuth metadata and a compatibility token flow for remote MCP. Set MCP_AUTH_SECRET in production so the MCP endpoints are not publicly open.

Project layout

src/
  index.ts   # Express app, MCP tools, SSE, OAuth helpers
  db.ts      # Mongoose connection
  models.ts  # ContentDraft, ResearchLog, AnalyticsEntry

Security

  • Store API keys, MongoDB URIs, and MCP_AUTH_SECRET only in .env or your host's secret manager.
  • Add a LICENSE file when you publish the repository.
Quick Setup
Installation guide for this server

Install Package (if required)

npx @modelcontextprotocol/server-ai-marketing-os-mcp

Cursor configuration (mcp.json)

{ "mcpServers": { "thebuildercompany-ai-marketing-os-mcp": { "command": "npx", "args": [ "thebuildercompany-ai-marketing-os-mcp" ] } } }