Zero to 33 Tools: Building the First MCP Server for ActiveCampaign

I searched GitHub, npm, PyPI, and every MCP registry I could find for an ActiveCampaign MCP server. Zero results. Not a bad one, not an incomplete one — nothing. For a platform with 185,000 paying customers and a full-featured API, that gap felt worth filling. So I built it from scratch: 33 tools covering contacts, deals, automations, tags, pipelines, campaigns, custom fields, and webhooks. What It Does MCP lets AI assistants interact with external tools directly. With this server installed, you manage your CRM and marketing automation without leaving your AI assistant. Before (manual): Log into ActiveCampaign Navigate to Contacts → search for a customer Check their tags, deals, automation history Switch to your AI tool Describe what you found Ask for analysis After (with mcp-activecampaign): "Find all contacts tagged 'enterprise-lead' and show me their deal pipeline status. Which ones haven't been contacted in 30 days?" Claude calls list_contacts with a tag filter, then list_deals for