WeCom MCP

WeCom MCP

Enables AI assistants to interact with Enterprise WeChat (WeCom) using natural language to manage meetings, book rooms, and search contacts. It supports multiple AI models and provides intelligent contact matching via Pinyin.

Category
Visit Server

README

WeCom MCP - 企业微信 AI 助手

基于 Laravel + MCP 协议构建的企业微信 AI 助手。用户通过自然语言指令,由 AI 自动识别意图并调用企业微信 API 完成操作。

"帮我明天下午3点和张三开个需求评审会" → AI 自动搜索联系人 → 创建会议 → 返回会议信息

功能特性

  • 自然语言驱动 — 用口语化指令完成企微操作,无需学习 API
  • MCP Tool 架构 — 基于 Model Context Protocol,可扩展的工具体系
  • 多模型支持 — Ollama / Claude / OpenAI / DeepSeek 一键切换
  • 拼音智能匹配 — 支持同音字、首字母缩写匹配联系人("王伟" = "汪伟")
  • 用户记忆 — 自动记住用户偏好和习惯,个性化响应
  • Claude Code Skills — 内置 API 文档查询 Skill,辅助开发

已支持的 MCP Tools

分类 Tool 说明
会议 CreateMeetingTool 创建在线会议
QueryMeetingsTool 查询会议列表
GetMeetingInfoTool 获取会议详情
UpdateMeetingTool 修改会议
CancelMeetingTool 取消会议
会议室 QueryMeetingRoomsTool 查询会议室列表
BookMeetingRoomTool 预定会议室
QueryRoomBookingsTool 查询预定记录
CancelRoomBookingTool 取消预定
联系人 SearchContactsTool 搜索联系人(拼音四级匹配)
记忆 SaveMemoryTool 保存用户偏好/习惯
DeleteMemoryTool 删除记忆

环境要求

  • PHP >= 8.2
  • Composer
  • SQLite(默认)或 MySQL
  • AI 模型(任选其一):
    • Ollama 本地部署(推荐开发环境,默认 qwen3:8b)
    • Claude API Key(Anthropic
    • OpenAI API Key
    • DeepSeek API Key

快速开始

1. 安装

git clone https://github.com/bangbangda/wecommcp.git
cd wecom-mcp
composer install
cp .env.example .env
php artisan key:generate
touch database/database.sqlite
php artisan migrate

2. 配置 AI 模型

编辑 .env,选择一个 AI 驱动:

# 方式一:Ollama 本地模型(无需 API Key)
AI_DRIVER=ollama
OLLAMA_BASE_URL=http://localhost:11434/v1
OLLAMA_MODEL=qwen3:8b

# 方式二:Claude
AI_DRIVER=anthropic
ANTHROPIC_API_KEY=sk-ant-xxx

# 方式三:OpenAI
AI_DRIVER=openai
OPENAI_API_KEY=sk-xxx

# 方式四:DeepSeek
AI_DRIVER=deepseek
DEEPSEEK_API_KEY=sk-xxx

3. 配置企业微信

企业微信管理后台创建自建应用,获取以下信息填入 .env

WECOM_CORP_ID=your_corp_id
WECOM_AGENT_SECRET=your_agent_secret
WECOM_AGENT_ID=your_agent_id
WECOM_AGENT_TOKEN=your_callback_token
WECOM_AGENT_AES_KEY=your_callback_aes_key

# 通讯录同步(可选)
WECOM_CONTACT_SECRET=your_contact_secret

4. 同步通讯录

php artisan wecom:sync-contacts

5. 开始对话

php artisan chat

输入自然语言指令即可:

You: 帮我明天下午两点和张三开个项目评审会
AI: 我来帮你创建会议。先搜索一下"张三"...
    找到联系人张三(产品部)。
    已创建会议「项目评审会」:
    - 时间:明天 14:00-15:00
    - 参会人:张三
    - 会议链接:https://meeting.tencent.com/xxx

架构概览

用户(语音/文字)
    |
    v
+-------------+     +------------------+     +-----------------+
|   接入层     | --> |   AI 编排层       | --> |  MCP Tool 执行层 |
|  企微回调    |     |  AiDriver        |     |  企微 API 调用   |
|  CLI 入口    | <-- |  + Tool Use      | <-- |  结果返回        |
+-------------+     +------------------+     +-----------------+
                            |
                    +-------+-------+
                    |       |       |
                  Ollama  Claude  OpenAI ...

核心设计原则:

  • Tool 参数隔离 — Tool 只暴露业务参数(title、start_time),access_token、corp_id 等基础设施参数由 Service 层内部管理,LLM 不可见
  • AI 驱动抽象 — Manager 模式,内部统一为 Claude 风格消息格式,Driver 负责格式转换
  • 拼音四级匹配 — 精确 → 拼音全匹配 → 首字母 → 模糊,多候选时 AI 自动追问

项目结构

app/
├── Ai/                         # AI 驱动抽象层
│   ├── Contracts/AiDriver.php  # 驱动接口
│   ├── Drivers/                # AnthropicDriver, OpenAiCompatibleDriver
│   ├── Dto/                    # AiResponse, ToolCall
│   └── AiManager.php           # Laravel Manager 工厂
├── Mcp/
│   ├── Servers/WecomServer.php # MCP Server 入口
│   └── Tools/                  # 12 个 MCP Tool
├── Models/                     # Contact, RecentMeeting, UserMemory
├── Services/
│   ├── ChatService.php         # AI 对话编排(tool 循环)
│   ├── ContactsService.php     # 拼音四级匹配
│   └── WecomService.php        # 企微 API 封装
└── Console/Commands/           # CLI 命令
config/ai.php                   # AI 驱动配置
skills/                         # Claude Code Skills

常用命令

php artisan chat                      # 交互式对话
php artisan test                      # 运行测试
php artisan wecom:sync-contacts       # 同步企微通讯录
php artisan mcp:start wecom           # 启动 MCP Server

# 切换 AI 驱动
AI_DRIVER=anthropic php artisan chat
AI_DRIVER=deepseek php artisan chat

添加新的 AI 驱动

  1. config/ai.phpdrivers 中添加配置
  2. 如果使用 OpenAI 兼容 API:在 AiManager 中添加 createXxxDriver() 返回 OpenAiCompatibleDriver
  3. 如果 API 格式不同:创建新 Driver 实现 AiDriver 接口
  4. .env 中添加对应环境变量

开发路线图

  • [x] 第一期:MVP — MCP Server + Tool + AI 驱动抽象层 + CLI 对话
  • [ ] 第二期:接入企微 — 回调接口、异步队列、多轮对话上下文(进行中)
  • [ ] 第三期:功能扩展 — 智能排期、部门级邀请、用户偏好挖掘
  • [ ] 第四期:架构升级 — MCP HTTP 模式部署、接入 Claude Desktop

技术栈

License

MIT

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured