Flowise

mcp-flowise is a Python package that implements a Model Context Protocol (MCP) server integrated with the Flowise API. It supports both dynamic tool registration and static configuration modes and is suitable for the integration of Flowise chat flows or assistants.

Artificial intelligence chatbots Developer tools #Flowise integration #Dynamic tools #MCP protocol #Chat flows .Python

rating : 2.5 points

downloads : 44

update time : 2025-04-28

What is mcp-flowise?

mcp-flowise is a Python tool that connects to the Flowise API. It allows you to manage and use Flowise's chat processes through the standardized Model Context Protocol (MCP). It simplifies the interaction with Flowise and provides dynamic tool registration and prediction creation functions.

How to use mcp-flowise?

You can install it automatically via Smithery or manually using the uvx package manager. After installation, configure the necessary environment variables to start using it.

Applicable scenarios

Suitable for scenarios where you need to integrate Flowise chat processes into existing MCP workflows, especially applications that require dynamic management of multiple chat processes or assistants.

Main features

Dynamic tool registrationAutomatically create dedicated tools for each Flowise chat process to simplify the integration process

Dual-mode operationSupports both LowLevel (dynamic) and FastMCP (static) modes to meet different complexity requirements

Flexible filtering mechanismSupports filtering chat processes through ID or name (regular expression) whitelist/blacklist

Advantages and limitations

Advantages

Seamless integration into the MCP ecosystem

Dynamic tool registration reduces manual configuration

Flexible filtering mechanism facilitates the management of a large number of chat processes

Supports Windows and Unix-like systems

Limitations

Requires Python 3.12 or higher

uvx installation on Windows may require additional configuration

Assistant functions are not fully implemented (on the TODO list)

How to use

Installation

Install automatically via Smithery or manually using uvx

Configuration

Set the necessary environment variables, especially FLOWISE_API_KEY and FLOWISE_API_ENDPOINT

Running

Start the server and integrate it into the MCP ecosystem

Usage examples

Usage in dynamic modeIn LowLevel mode, each chat process automatically becomes an available tool

Usage in static modeIn FastMCP mode, use standard tools to interact with the specified chat process

Frequently Asked Questions

How to choose the operation mode?

What should I do if I encounter problems during Windows installation?

How to filter the displayed chat processes?

Related resources

GitHub repository

Project source code and latest updates

Smithery installation page

Install automatically via Smithery

Flowise official website

Flowise official documentation

🚀 mcp-flowise

mcp-flowise is a Python package that implements a Model Context Protocol (MCP) server and integrates with the Flowise API. This package is used to register and manage chatflows on the Flowise platform.

🚀 Quick Start

Start the MCP server:

mcp-flowise --port 8000

By default, the server runs on port 8000.

✨ Features

1. Register Chatflows

from mcp_flowise import MCPFlowise

client = MCPFlowise(
    api_key="your_api_key",
    api_endpoint="http://localhost:3000"
)

response = client.register_chatflow(name="My Chatflow", description="A test chatflow")
print(response)

2. List Chatflows

Get the list of registered chatflows:

from mcp_flowise import MCPFlowise

client = MCPFlowise(
    api_key="your_api_key",
    api_endpoint="http://localhost:3000"
)

chatflows = client.list_chatflows()
print(chatflows)

3. Delete Chatflows

Delete a specified chatflow:

from mcp_flowise import MCPFlowise

client = MCPFlowise(
    api_key="your_api_key",
    api_endpoint="http://localhost:3000"
)

response = client.delete_chatflow(chatflow_id="chatflow_id")
print(response)

📦 Installation

To install mcp-flowise, run the following command:

pip install mcp-flowise

📚 Documentation

Environment Variables

Required Configuration

FLOWISE_API_KEY: Flowise API token (required).
FLOWISE_API_ENDPOINT: The base URL of Flowise, defaulting to http://localhost:3000.

Additional Configuration

FLOWISE_CHATFLOW_DESCRIPTIONS: A comma-separated list of chatflow_id:description pairs, for example:
```
FLOWISE_CHATFLOW_DESCRIPTIONS="abc123:Chatflow One,xyz789:Chatflow Two"
```

Filtering Chatflows

You can use the following environment variables to filter chatflows:

Whitelist by ID:
FLOWISE_WHITELIST_ID="id1,id2,id3"
Blacklist by ID:
FLOWISE_BLACKLIST_ID="id4,id5"
Whitelist by Name Regex:
FLOWISE_WHITELIST_NAME_REGEX=".*important.*"
Blacklist by Name Regex:
FLOWISE_BLACKLIST_NAME_REGEX=".*deprecated.*"

⚠️ Important Note

The whitelist takes precedence over the blacklist. If both the whitelist and blacklist are set, the strictest rule will apply.

Security

Protect API Keys: Ensure that FLOWISE_API_KEY is not leaked or exposed in logs or code repositories.
Environment Configuration: Use the .env file or environment variables to manage sensitive configurations.

Add .env to your .gitignore:

# .gitignore
.env

Troubleshooting

Missing API Key: Make sure FLOWISE_API_KEY is set correctly.
Invalid Configuration: If both FLOWISE_CHATFLOW_ID and FLOWISE_ASSISTANT_ID are set, the server will refuse to start.
Connection Error: Verify that FLOWISE_API_ENDPOINT is reachable.