Skip to main content
The MCP Server exposes AI Flows as callable tools that can be consumed by AI agents and external systems using the Model Context Protocol (MCP). It provides a standardized interface for discovering, invoking, and governing executable capabilities across a team. An MCP Tool represents a published AI Flow that has been enabled for external or agent-based invocation. MCP tools expose defined inputs, outputs, and execution behavior, allowing agents and systems to reliably execute workflows, retrieve data, and perform actions.

​
Overview

All AI Flows created across all GIDRs in a team are listed on the MCP Server page. Any AI Flow can be enabled or disabled as an MCP Tool.
  • When an AI Flow is enabled as an MCP Tool, it becomes callable through the MCP Server.
  • When disabled, the flow remains available internally but cannot be invoked through MCP.
  • This allows teams to safely control which capabilities are exposed for integration and automation.
Once one or more tools are enabled, the MCP Server URL becomes available. This endpoint can be copied and used by:
  • External MCP clients integrating with GIDR to invoke tools programmatically
  • ReAct agents within GIDR’s Agentic AI Studio to dynamically discover and call tools during reasoning and execution
This creates a unified tool interface for both internal agent orchestration and external system integration. MCP Server Overview