enterprise-knowledge-mcp
Enterprise Knowledge MCP
Ein Enterprise-Wissensmanagementdienst auf Basis des Model Context Protocol (MCP).
Der Dienst bietet drei Kernfunktionen für große Sprachmodelle:
Werkzeug | Funktion |
| Semantische Suche in der Unternehmenswissensbasis |
| Vollständiges Dokument anhand der Dokument-ID abrufen |
| Unternehmensprojektfälle durchsuchen |
Projektstruktur
enterprise-knowledge-mcp/
├── src/
│ └── enterprise_knowledge_mcp/
│ ├── __init__.py # 包入口
│ ├── server.py # MCP 工具注册(协议层)
│ └── retriever.py # 检索逻辑(业务层)
├── tests/
│ ├── test_server.py # 单元测试
│ └── test_e2e.py # MCP Client 端到端测试
├── scripts/
│ ├── seed_data.py # 导入示例数据到 ChromaDB
│ └── download_model.py # 下载嵌入模型(离线环境用)
├── pyproject.toml
├── server.json # MCP 服务描述清单
├── README.md
└── LICENSERelated MCP server: MCP Knowledge Base Server
Architekturdesign
┌─────────────────────────────────────────┐
│ server.py (协议层) │
│ - 注册 MCP Tool │
│ - 参数校验 & 输出格式化 │
└──────────────┬──────────────────────────┘
│ 调用
┌──────────────▼──────────────────────────┐
│ retriever.py (业务层) │
│ - KnowledgeRetriever 知识库检索 │
│ - DocumentRetriever 文档检索 │
│ - CaseRetriever 案例检索 │
└──────────────┬──────────────────────────┘
│ 替换实现
┌──────────────▼──────────────────────────┐
│ 数据层(可插拔) │
│ - MockRetriever 开发/测试用 │
│ - ChromaRetriever ChromaDB 向量检索 │
│ - 自定义 Retriever ES / 数据库 / ... │
└─────────────────────────────────────────┘Der Retriever verwendet ein austauschbares Design aus Basisklasse + Implementierung. In server.py muss nur das instanziierte Objekt ausgetauscht werden, um das Backend zu wechseln; der Tool-Registrierungscode muss nicht geändert werden.
Schnellstart
Installation
# 克隆仓库
git clone https://github.com/your-username/enterprise-knowledge-mcp.git
cd enterprise-knowledge-mcp
# 创建虚拟环境
python -m venv .venv
.venv\Scripts\activate # Windows
# source .venv/bin/activate # Linux/macOS
# 安装(开发模式)
pip install -e ".[dev]"ChromaDB-Daten initialisieren
# 导入示例企业知识数据(知识库 5 条、文档 3 篇、案例 4 个)
python scripts/seed_data.pyDienst starten
# Mock 模式(无需 ChromaDB 数据)
python -m enterprise_knowledge_mcp.server
# ChromaDB 模式(需先执行 seed_data.py)
RETRIEVER_BACKEND=chroma python -m enterprise_knowledge_mcp.serverTests ausführen
pytest tests/ -vAnbindung von MCP-Clients
Claude Desktop
Fügen Sie in claude_desktop_config.json Folgendes hinzu:
{
"mcpServers": {
"enterprise-knowledge": {
"command": "python",
"args": ["-m", "enterprise_knowledge_mcp.server"],
"cwd": "C:/path/to/enterprise-knowledge-mcp"
}
}
}Python SDK
from mcp import ClientSession, StdioServerParameters
from mcp.client.stdio import stdio_client
server_params = StdioServerParameters(
command="python",
args=["-m", "enterprise_knowledge_mcp.server"],
)
async with stdio_client(server_params) as (read, write):
async with ClientSession(read, write) as session:
await session.initialize()
result = await session.call_tool(
"search_knowledge", {"query": "人工智能"}
)
print(result.content[0].text)Benutzerdefinierter Retriever
Leiten Sie von der Basisklasse ab und implementieren Sie die Methoden retrieve / get_document / search_cases:
from enterprise_knowledge_mcp.retriever import (
KnowledgeRetriever,
SearchResult,
)
class MyRetriever(KnowledgeRetriever):
def retrieve(self, query: str, top_k: int = 5) -> list[SearchResult]:
# 你的检索逻辑
return [SearchResult(text="...", source="...", score=0.9)]Nehmen Sie anschließend in server.py den Austausch vor:
from .retriever import MyRetriever
knowledge_retriever = MyRetriever()Lizenz
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Search and manage your team's knowledge from MCP clients.
Make your knowledge agent-ready. One MCP endpoint, 5 connectors, 3 search modes.
Knowledge base MCP for AI agents on iknow.dev. Search, read, and maintain via OAuth.
Knowledge coverage map and health score. Ingest docs into a governed knowledge graph via MCP.
Related MCP Servers
- AlicenseBqualityDmaintenanceEnables interaction with the Innovaas Knowledge Management System through MCP, providing multi-modal search, RAG-powered chat with intelligent token management, and document access.7MIT
- FlicenseCqualityDmaintenanceEnables semantic search over knowledge-base articles and listing of sample support tickets using MCP tools.4-
- FlicenseNot gradedqualityBmaintenanceEnables to build and query a knowledge base with retrieval-augmented generation, supporting document ingestion, hybrid search, and live data integration from external APIs via MCP tools.2-
- FlicenseNot gradedqualityCmaintenanceEnables AI agents and users to search a unified organizational knowledge warehouse, create and iterate on documents, manage folders and reviews, and curate shareable knowledge packs through any compatible MCP client.-