Home / MCP / Flowise MCP Server

Flowise MCP Server

Exposes Flowise chatflows as MCP tools to list chatflows, create predictions, and dynamically register tools

python

Installation

Add the following to your MCP client configuration file.

Configuration

{
    "mcpServers": {
        "flowise": {
            "command": "C:\\Users\\matth\\.local\\bin\\uvx.exe",
            "args": [
                "--from",
                "C:\\Users\\matth\\downloads\\mcp-flowise",
                "mcp-flowise"
            ],
            "env": {
                "FLOWISE_API_KEY": "${FLOWISE_API_KEY}",
                "FLOWISE_API_ENDPOINT": "${FLOWISE_API_ENDPOINT}",
                "LOGLEVEL": "ERROR",
                "APPDATA": "C:\\Users\\matth\\AppData\\Roaming"
            }
        }
    }
}

You deploy mcp-flowise to connect Flowise chatflows with MCP workflows. It exposes tools for listing chatflows, creating predictions, and dynamically registering chatflow tools to streamline interactions within your MCP environment.

How to use

Start by running mcp-flowise as a local MCP server. Depending on your setup, you can use a simple, static mode or a dynamic mode that exposes tools for every Flowise chatflow. Once running, configure your MCP client to connect to the server and begin listing chatflows, creating predictions, and invoking chatflow tools in your workflows.

How to install

Prerequisites: Python 3.12 or higher and the uvx package manager. You will also need access to the Flowise API to enable the MCP integration.

Step-by-step commands to run mcp-flowise locally or in your MCP ecosystem are shown below. Follow the commands exactly as written to ensure proper startup and configuration.

{
  "mcpServers": {
    "flowise": {
      "command": "C:\\Users\\matth\\.local\\bin\\uvx.exe",
      "args": [
        "--from",
        "C:\\Users\\matth\\downloads\\mcp-flowise",
        "mcp-flowise"
      ],
      "env": {
        "LOGLEVEL": "ERROR",
        "APPDATA": "C:\\Users\\matth\\AppData\\Roaming",
        "FLOWISE_API_KEY": "your-api-key-goes-here",
        "FLOWISE_API_ENDPOINT": "http://localhost:3010/"
      }
    }
  }
}

Additional configuration and notes

Environment variables control access and behavior. Provide the Flowise API key and endpoint, and adjust log levels as needed. If you need to adapt to Windows path conventions, use the exact paths shown in the example or replace them with your own valid paths while preserving the structure.

Troubleshooting

Missing or invalid API key: ensure FLOWISE_API_KEY is set to a valid key. Connection errors: verify FLOWISE_API_ENDPOINT is reachable from your runtime environment. If the server refuses to start due to conflicting settings, double-check that only one mode is configured at a time (either FastMCP or LowLevel).

Security and best practices

Keep API keys private. Store sensitive configuration in environment variables or secure secret management, and avoid logging them. Use .env files or your platform’s secret management features and add .env to your ignore list.

Available tools

list_chatflows

Lists all available chatflows in Flowise so you can see what you can invoke or create predictions against.

create_prediction

Create a prediction for a specified chatflow or assistant, returning results from the Flowise backend.

dynamic_chatflow_tool

Dynamically exposed tool for a chatflow when LowLevel mode is enabled; tool name matches the chatflow and description may be provided via FLOWISE_CHATFLOW_DESCRIPTIONS.