VerticalAI docs
Guides

MCPs

Connect external systems to your workspace through MCP connectors — the catalogue, your own connectors, and who can manage them.

MCPs are connectors that let your agent reach external systems through the Model Context Protocol. A connector exposes a set of capabilities the agent can use, without you hand-writing a webhook tool for every endpoint. They are managed at the workspace level, on the MCPs page.

Catalogue and My MCPs

The MCPs page has two tabs:

  • Catalogue — connectors you can add, ready to connect to your workspace.
  • My MCPs — the connectors your workspace has added, with their build logs and a deployment timeline so you can see their status over time.

Add one from the catalogue and it appears under My MCPs.

Who can do what

Access follows your workspace role:

  • Owner, admin, and editor can add connectors.
  • Owner and admin can manage them — configure and remove.

Each connector also tracks ownership, so contributors act within their role.

How MCPs relate to tools

Tools are the agent's per-call actions; an MCP connector is a managed integration that backs a set of capabilities. Use a webhook tool for a single endpoint you own; reach for an MCP connector when you want to plug in a whole external system that already speaks MCP, rather than wrapping each of its endpoints by hand.

On this page