Skip to main content
Activepieces includes a built-in MCP server that allows AI assistants (Claude, GPT, etc.) to interact with your automation platform. AI agents can build flows, manage tables, test automations, and more — all through natural language.

​
What is MCP?

The Model Context Protocol is an open standard that lets AI assistants connect to external tools and data sources. Activepieces implements an MCP server that exposes 30 tools for managing your automation workflows.

​
Enable the MCP Server

  1. Go to Settings → MCP Server
  2. Toggle the server on
  3. Copy the server URL and token

​
Connect Your AI Assistant

The MCP server uses the Streamable HTTP transport. Configure your AI client with:
  • URL: https://your-instance.com/api/v1/projects/{projectId}/mcp-server/http
  • Authentication: Bearer token (or ?token= query parameter)
The token can be rotated at any time from the MCP Server settings page.

​
Claude Desktop

Add this to your Claude Desktop config (claude_desktop_config.json): 
{
  "mcpServers": {
    "activepieces": {
      "url": "https://your-instance.com/api/v1/projects/{projectId}/mcp-server/http?token={your-token}"
    }
  }
}

​
Cursor / Other MCP Clients

Use the Streamable HTTP URL with your token as a Bearer token in the Authorization header.

​
Tool Categories

Tools are organized into categories. Discovery tools are always available. Other categories can be enabled or disabled per-project in the MCP Server settings.
CategoryToolsDescription
Discovery10Read-only tools for exploring flows, pieces, tables, runs, and AI models
Flow Management4Create, rename, publish, and enable/disable flows
Flow Building4Add, update, and delete steps and triggers
Router & Branching2Add and delete conditional branches
Annotations1Manage canvas notes
Tables6Full CRUD for tables, fields, and records
Testing & Runs3Test flows, inspect results, retry failures
See the Tools Reference for the complete catalog with input schemas and examples.

​
Security

  • Credentials are never exposed — connection secrets, API keys, and OAuth tokens are never returned by any MCP tool
  • Project-scoped — all operations are scoped to the authenticated project
  • Sensitive setup — the ap_setup_guide tool returns instructions for the user to configure connections and AI providers in the UI, rather than handling secrets through MCP
  • Token rotation — the MCP token can be rotated at any time without affecting existing connections or flows