Skip to content
Stories & Systems
← Case studies
AI Infrastructure

461 Tools Where There Used to Be 36

CRM work that used to need bespoke Make.com flows now resolves through plain conversation — 461 agent tools where the official server offered 36.

461 Tools Where There Used to Be 36
Industry
Developer Tooling / AI
Engagement
Internal R&D, productized for clients
Timeline
Multi-month build, continuous expansion
Built with
  • Node.js
  • TypeScript
  • Model Context Protocol
  • GoHighLevel API
  • OpenAPI

The outcome

  • Client engagements that previously needed bespoke Make.com flows now resolve through agentic conversation.
  • Every category the official MCP missed — blogs, social media, granular opportunity updates, custom fields — became accessible.
  • It sets up a moat: any agency relying on the official MCP is structurally limited in what it can offer.

The challenge

The official GoHighLevel MCP server exposes 36 tools — enough for trivial automations, but not for real agentic work across the platform's 19 functional categories. Every Stories and Systems engagement that wanted AI-driven CRM work ran into the same ceiling.

How we did it

  • Reverse-mapped the entire GoHighLevel API surface into a modular MCP architecture organized by category: contacts, opportunities, calendars, conversations, blogs, social posting, payments, locations, and more.
  • Built 461-plus tools, each with strict input validation, opinionated defaults, and AI-readable descriptions tuned for high triggering accuracy.
  • Category modules load selectively so the token budget stays controlled.
  • Productized to deploy per client with their own credentials, isolated from other tenants.

The official MCP felt like being handed a screwdriver to disassemble an engine. So we built the rest of the toolbox.

Stories & Systems build log

Have a bottleneck that needs a system?

Find the leak in five minutes, and the first three things to fix.

Start the Friction Assessment