MCP Servers

Connect custom MCP (Model Context Protocol) servers to extend your agents' capabilities beyond built-in integrations.

MCP servers let you connect any tool or service that supports the Model Context Protocol to your agents. This extends your team's capabilities beyond the built-in integrations to anything you can expose through an MCP interface.

What is MCP?

The Model Context Protocol is a standard for connecting AI agents to external tools and data sources. If a service exposes an MCP server, your Vibey agents can use it - browse available tools, call them, and work with the results.

Adding an MCP Server

Open Workspace Settings

Navigate to Workspace Settings → MCP Servers.

Add a server

You can add servers in two ways:

Form mode - Enter the server name, URL, optional description, and domain assignment one at a time.

JSON mode - Paste a JSON array to import multiple servers at once.

Assign a domain

Choose which agents can access this server based on their domain:

Domain	Who gets access
Shared	All agents
Marketing	Marketing-domain agents only
Analyst	Analyst-domain agents only
Developer	Developer-domain agents only

Enable for agents

Toggle Agent Enabled to make the server's tools available to your agents during missions and Team conversations.

Server Configuration

Each MCP server has:

Field	Description
Name	Display name for the server
Server URL	The MCP endpoint URL
Description	Optional description of what the server provides
Domain	Which agent domain gets access (shared, marketing, analyst, developer)
API Key	Optional authentication key
Headers	Optional custom HTTP headers
Enabled	Whether the server is active
Agent Enabled	Whether agents can use the server's tools

Managing Servers

From the MCP settings page you can:

See all connected servers with their tool count and last connection time
Browse cached tools for each server (name and description)
Toggle servers on/off
Toggle agent access per server
Edit server configuration
Delete servers

JSON Bulk Import

For adding multiple servers at once, switch to JSON mode and paste an array:

[
  {
    "name": "My Custom Tool",
    "url": "https://mcp.example.com",
    "description": "Custom data processing tools",
    "domain": "analyst",
    "agent_enabled": true
  }
]

Each entry supports: name, url, description, domain, api_key, headers, enabled, agent_enabled.

Use Cases

What you want	MCP server to connect
Custom internal APIs	Your company's API wrapped in an MCP server
Specialized data sources	Industry databases, research tools
Custom automation	Workflow tools, internal systems
Third-party tools not in the built-in library	Any service with MCP support

MCP servers are the escape hatch for when the built-in integrations don't cover what you need. If a tool has an MCP server, your agents can use it.