FlowDot MCP Server

The FlowDot MCP Server (@flowdot.ai/mcp-server) provides 200+ tools, plus 8 educational learn:// resources, for integrating FlowDot with AI clients like Claude Desktop, Cursor IDE, Claude Code, and any MCP-compatible application.

What is MCP?

The Model Context Protocol (MCP) is an open standard for connecting AI assistants to external tools and data sources. It allows AI clients like Claude Desktop to access capabilities beyond their built-in features.

FlowDot's MCP server implements the full MCP specification, providing tools that let AI assistants:

• Create, edit, and execute workflows
• Build and publish apps
• Manage custom nodes and toolkits
• Query knowledge bases with semantic search
• Design recipes (multi-step agentic workflows)

Installation

Run the MCP server using npx (no global installation needed):

npx "@flowdot.ai/mcp-server"

Get Your Token

1. Go to flowdot.ai and log in
2. Navigate to Settings > MCP Tokens
3. Click "Create New Token"
4. Select the scopes you need (or use full access)
5. Copy the token (starts with fd_mcp_)

Integration Points

The FlowDot MCP server works with any MCP-compatible AI client:

200+ Tools

The MCP server exposes a complete operational interface to the FlowDot platform across two dozen categories — no web UI required. Every tool is scope-gated (see the MCP Token System below).

Learn Resources (learn://)

The server also exposes 8 standalone concept guides via the MCP resource interface. Read the one that matches your task before invoking tools in that area — each is a complete guide with concepts, tool reference, and worked examples.

• learn://overview — FlowDot platform overview (start here)
• learn://workflows — Workflow creation guide
• learn://recipes — Agent recipe orchestration guide
• learn://custom-nodes — Custom node development guide
• learn://apps — App development guide
• learn://toolkits — Agent toolkit guide
• learn://knowledge-base — Knowledge base & RAG guide
• learn://characters — Agent character (voice call) setup guide

MCP Token System

FlowDot uses granular scopes to control what each MCP token can access.

25 Granular Scopes

When creating a token, you select exactly which scopes to grant. Restrict each token to the minimum scope it needs:

• workflows:read / workflows:execute / workflows:manage - View, run, and build workflows
• executions:read / executions:manage - Status, history, stream; cancel, retry
• agent:chat - Agent chat
• custom_nodes:read / custom_nodes:manage - Custom node operations
• apps:read / apps:manage - App management, code and file editing
• recipes:read / recipes:manage - Recipe design, steps, and stores
• agent_characters:read / agent_characters:manage - Voice-call character operations
• knowledge:read / knowledge:manage - Knowledge base and RAG
• agent_toolkits:read / agent_toolkits:manage - Toolkit operations and invocation
• sharing:read / sharing:manage - Shared results, public URLs, votes, comments
• input_presets:read / input_presets:manage - Input preset operations
• teams:read - List teams
• discovery:read - Search, tags, public browsing
• analytics:read - Metrics, comments, history

Token Management

Manage tokens at Settings > MCP Tokens:

• Create tokens with specific scopes
• View active tokens
• Revoke tokens at any time
• See last used timestamp

Configuration Examples

Claude Desktop

Add to your Claude Desktop config — macOS: ~/Library/Application Support/Claude/claude_desktop_config.json, Windows: %APPDATA%\Claude\claude_desktop_config.json:

{
  "mcpServers": {
    "flowdot": {
      "command": "npx",
      "args": ["@flowdot.ai/mcp-server"],
      "env": {
        "FLOWDOT_API_TOKEN": "fd_mcp_your_token_here"
      }
    }
  }
}

Cursor IDE

Add to your Cursor MCP settings:

{
  "mcpServers": {
    "flowdot": {
      "command": "npx",
      "args": ["@flowdot.ai/mcp-server"],
      "env": {
        "FLOWDOT_API_TOKEN": "fd_mcp_your_token_here"
      }
    }
  }
}

Environment Variables

Alternatively, set the token as an environment variable:

# macOS/Linux
export FLOWDOT_API_TOKEN=fd_mcp_your_token_here

# Windows
set FLOWDOT_API_TOKEN=fd_mcp_your_token_here

Related Documentation