- Explore MCP Servers
- mcp-server-hub
Mcp Server Hub
What is Mcp Server Hub
mcp-server-hub is a collection of tools designed for the Model Context Protocol (MCP), enabling the invocation of various AI service APIs, including natural language processing (NLP), computer vision (CV), and speech processing tasks.
Use cases
Use cases for mcp-server-hub include building conversational agents, generating text based on prompts, classifying images, summarizing long texts, and recognizing speech in audio files. It can be utilized in applications ranging from customer support bots to creative writing tools.
How to use
To use mcp-server-hub, first install the necessary dependencies with ‘npm install’. Then, configure your API keys in a ‘.env’ file. You can start the Dify MCP service with ‘npm run start:dify’ and utilize the Hugging Face MCP service by importing the relevant functions and calling them with appropriate parameters.
Key features
Key features of mcp-server-hub include support for text generation, classification, question answering, summarization, image classification and segmentation, speech recognition, and zero-shot learning tasks. It also includes Dify services for managing conversational applications.
Where to use
mcp-server-hub can be used in various fields such as AI development, chatbot creation, content generation, image processing, and speech recognition applications, making it suitable for developers and researchers in the AI domain.
Clients Supporting MCP
The following are the main client software that supports the Model Context Protocol. Click the link to visit the official website for more information.
Overview
What is Mcp Server Hub
mcp-server-hub is a collection of tools designed for the Model Context Protocol (MCP), enabling the invocation of various AI service APIs, including natural language processing (NLP), computer vision (CV), and speech processing tasks.
Use cases
Use cases for mcp-server-hub include building conversational agents, generating text based on prompts, classifying images, summarizing long texts, and recognizing speech in audio files. It can be utilized in applications ranging from customer support bots to creative writing tools.
How to use
To use mcp-server-hub, first install the necessary dependencies with ‘npm install’. Then, configure your API keys in a ‘.env’ file. You can start the Dify MCP service with ‘npm run start:dify’ and utilize the Hugging Face MCP service by importing the relevant functions and calling them with appropriate parameters.
Key features
Key features of mcp-server-hub include support for text generation, classification, question answering, summarization, image classification and segmentation, speech recognition, and zero-shot learning tasks. It also includes Dify services for managing conversational applications.
Where to use
mcp-server-hub can be used in various fields such as AI development, chatbot creation, content generation, image processing, and speech recognition applications, making it suitable for developers and researchers in the AI domain.
Clients Supporting MCP
The following are the main client software that supports the Model Context Protocol. Click the link to visit the official website for more information.
Content
MCP 服务工具
这是一个 MCP(Model Context Protocol)服务工具集合,用于调用各种 AI 服务 API。
功能特点
Hugging Face MCP 服务
- 支持文本生成、分类、问答、摘要等 NLP 任务
- 支持图像分类、分割、图像到文本等 CV 任务
- 支持语音识别等语音处理任务
- 支持零样本学习分类任务
Dify MCP 服务 (新增)
- 支持对话型应用 API 调用
- 支持文本生成型应用 API 调用
- 支持对话历史管理
- 支持对话列表查询和管理
安装
首先,安装必要的依赖:
npm install
配置
在项目根目录创建 .env 文件,并添加以下配置:
# Hugging Face API 密钥(如果使用 Hugging Face 服务) HUGGINGFACE_API_KEY=your_huggingface_api_key # Dify API 配置(如果使用 Dify 服务) DIFY_API_KEY=your_dify_api_key DIFY_BASE_URL=your_dify_base_url
使用方法
启动 Dify MCP 服务
npm run start:dify
Hugging Face MCP 服务
import { createHuggingFaceMcpTool } from './src/hf_mcp';
// 创建工具实例
const hfTool = createHuggingFaceMcpTool('your_huggingface_api_key');
// 使用工具调用模型
async function example() {
const result = await hfTool.textGeneration(
'gpt2',
'人工智能正在改变世界,',
{ max_new_tokens: 50 }
);
console.log(result);
}
example();
Dify MCP 服务功能
Dify MCP 服务提供以下工具:
-
chat_with_dify:向 Dify 发送聊天消息并获取回复
- 参数:
base_url:Dify API 基础 URLapi_key:Dify API 密钥query:用户查询内容inputs:输入参数(可选)conversation_id:对话 ID(可选,用于继续对话)user:用户标识(可选)
- 参数:
-
generate_with_dify:使用 Dify 生成文本内容
- 参数:
base_url:Dify API 基础 URLapi_key:Dify API 密钥inputs:输入参数user:用户标识(可选)
- 参数:
-
get_conversation_history:获取 Dify 对话历史记录
- 参数:
base_url:Dify API 基础 URLapi_key:Dify API 密钥conversation_id:对话 IDuser:用户标识(可选)
- 参数:
-
list_conversations:获取 Dify 对话列表
- 参数:
base_url:Dify API 基础 URLapi_key:Dify API 密钥user:用户标识(可选)first:返回结果数量(可选,默认 20)
- 参数:
-
rename_conversation:重命名 Dify 对话
- 参数:
base_url:Dify API 基础 URLapi_key:Dify API 密钥conversation_id:对话 IDname:新对话名称user:用户标识(可选)
- 参数:
Dify API 使用示例
使用 Dify 聊天功能
// 向 Dify 发送聊天消息
const response = await difyClient.chatMessage(
"你好,请介绍一下自己", // 查询内容
{}, // 输入参数
"conversation_id_123", // 对话 ID(可选)
"user_123" // 用户标识(可选)
);
console.log(response.answer); // 输出 Dify 的回复
使用 Dify 文本生成功能
// 使用 Dify 生成文本
const response = await difyClient.completionMessage(
{ prompt: "写一篇关于人工智能的短文" }, // 输入参数
"user_123" // 用户标识(可选)
);
console.log(response.answer); // 输出生成的文本
Hugging Face 支持的功能
-
文本生成
await hfTool.textGeneration('gpt2', '输入文本', { max_new_tokens: 50 }); -
文本分类
await hfTool.textClassification('distilbert-base-uncased-finetuned-sst-2-english', '这是一段文本'); -
问答
await hfTool.questionAnswering('deepset/roberta-base-squad2', { question: '问题?', context: '包含答案的上下文' }); -
填充遮罩
await hfTool.fillMask('bert-base-uncased', 'Paris is the [MASK] of France.'); -
摘要生成
await hfTool.summarization('facebook/bart-large-cnn', '长文本内容'); -
文本翻译
await hfTool.translation('t5-base', '要翻译的文本'); -
图像分类
await hfTool.imageClassification('google/vit-base-patch16-224', imageBlob); -
图像分割
await hfTool.imageSegmentation('facebook/detr-resnet-50-panoptic', imageBlob); -
图像到文本
await hfTool.imageToText('nlpconnect/vit-gpt2-image-captioning', imageBlob); -
零样本图像分类
await hfTool.zeroShotImageClassification( 'openai/clip-vit-base-patch32', imageBlob, ['猫', '狗', '鸟'] ); -
零样本文本分类
await hfTool.zeroShotClassification( 'facebook/bart-large-mnli', '这是一段文本', ['类别1', '类别2', '类别3'], false ); -
语音识别
await hfTool.automaticSpeechRecognition('facebook/wav2vec2-base-960h', audioBlob);
注意事项
- 使用前需要先获取相应的 API 密钥
- 请注意 API 调用限制和费用
- Dify API 的具体参数和返回值可能因应用配置不同而有所差异
许可证
MIT
Dev Tools Supporting MCP
The following are the main code editors that support the Model Context Protocol. Click the link to visit the official website for more information.










