​
MCP Server API Overview

Kubiya’s MCP server provides a comprehensive set of tools for AI assistants to interact with the Kubiya platform through the Model Context Protocol. This section documents the core API endpoints and their usage.

​
API Categories

​
Tool Execution

  • execute_tool: Execute container-based tools with live streaming
  • create_on_demand_tool: Create and execute tools from definitions
  • execute_whitelisted_tool: Execute pre-approved tools with security controls

​
Workflow Management

  • execute_workflow: Run complete multi-step workflows
  • list_workflows: Discover available workflows
  • get_workflow_status: Monitor workflow execution

​
Infrastructure Management

  • list_runners: List available execution runners
  • check_runner_health: Monitor runner health status
  • find_available_runner: Auto-select optimal runners

​
Agent Interaction

  • list_agents: Discover available AI agents
  • chat_with_agent: Engage in multi-turn conversations
  • execute_agent_task: Run agent-specific tasks

​
Authentication

All MCP server interactions require a valid Kubiya API key configured in the environment: 
export KUBIYA_API_KEY="kb-your-api-key"

​
Next Steps