Dify Workflow Server MCP Server

Developer ToolsGo

Execute and manage Dify workflows through a Model Context Protocol server

Available Tools

list_workflows

Lists all authorized Dify workflows available for execution

execute_workflow

Executes a specified Dify workflow with a custom input message

workflow_nameinput

Dify Workflow Server is a Model Context Protocol server that enables seamless integration with Dify's workflow capabilities. It allows you to query and invoke multiple custom Dify workflows on demand, providing a bridge between AI assistants and the powerful workflow automation features of the Dify platform. With this integration, AI assistants can execute complex workflows created in Dify directly through simple commands.

Overview

Dify Workflow Server provides a Model Context Protocol (MCP) interface to interact with workflows created on the Dify platform. This integration allows AI assistants to list available workflows and execute them with custom inputs, extending the capabilities of AI assistants with Dify's powerful workflow automation features.

Installation

Prerequisites

Go 1.x (if building from source)
Access to a Dify platform instance with created workflows
API keys for your Dify workflows

Installation Methods

Method 1: Build from source

# Clone the repository
git clone https://github.com/gotoolkits/mcp-difyworkflow-server.git

# Navigate to the directory
cd mcp-difyworkflow-server

# Build the application
go build .
# or use make
make build

Method 2: Use pre-built binary

Download the latest release from the GitHub releases page.

After installation, you may want to create a symbolic link for easier access:

sudo ln -s /path/to/mcp-difyworkflow-server /usr/local/bin/mcp-difyworkflow-server

Configuration

To use the Dify Workflow Server with your AI assistant, you need to configure it with your Dify platform details and API keys.

Add the following configuration to your AI assistant's MCP configuration:

{
  "mcpServers": {
    "mcp-difyworkflow-server": {
      "command": "mcp-difyworkflow-server",
      "args": ["-base-url", "http://your-dify-instance/v1"],
      "env": {
        "DIFY_WORKFLOW_NAME": "workflow-name1,workflow-name2",
        "DIFY_API_KEYS": "api-key1,api-key2"
      }
    }
  }
}

Configuration Parameters

base-url: The base URL of your Dify platform API server
DIFY_WORKFLOW_NAME: Comma-separated list of workflow names you want to make available
DIFY_API_KEYS: Comma-separated list of API keys corresponding to each workflow (must be in the same order as the workflow names)

Important Note: The input variable name in your Dify workflows should be defined as "message" for compatibility with this server.

Usage

Once configured, you can interact with your Dify workflows through the AI assistant using natural language. The server provides two main functions:

Listing available workflows
Executing a specific workflow with custom input

For example, you can ask:

"Show me the list of available Dify workflows"
"Execute the workflow named 'workflow-translator' with the message 'This is a test message'"

The server will handle these requests and return the results from the Dify platform.