Enterprise MCP Gateway and Tool Registry
Enables AI agents to discover and execute tools via a secure MCP server with JWT authentication, RBAC, rate limiting, and audit logging.
README
Enterprise MCP Gateway and AI Tool Registry
Secure tool-discovery and execution gateway for AI-agent integrations. The project combines a runnable FastAPI gateway with an optional standards-based MCP server entrypoint built with the official Python MCP SDK.
Highlights
- JWT authentication with demo admin, developer, and viewer roles
- RBAC enforcement for tool execution and administrative operations
- Persistent tool registry and audit logs with SQLAlchemy
- Built-in calculator, summarizer, keyword extractor, and gateway-health tools
- Allowlisted HTTP tool registration for controlled external API integration
- Redis-backed rate limiting with an in-memory fallback
- Optional OpenTelemetry export to Jaeger
- Prometheus-style gateway metrics at
/metrics - REST endpoints for browser testing and an optional FastMCP Streamable HTTP server
- Automated tests and GitHub Actions CI
Architecture
AI Client / Developer
|
v
FastAPI Gateway -----> JWT Auth + RBAC
|
+-----> Tool Registry -----> Built-in Tools
| `----> Allowlisted HTTP Tools
|
+-----> Rate Limiter ------> Redis or in-memory fallback
|
+-----> Audit Store -------> SQLite quick start / PostgreSQL Docker mode
|
`-----> OpenTelemetry -----> Jaeger (optional)
Quick Start
python -m venv .venv
source .venv/bin/activate
python -m pip install -r requirements-dev.txt
cp .env.example .env
python -m uvicorn app.main:app --host 0.0.0.0 --port 8000
Open http://localhost:8000/docs.
Demo Accounts
| Username | Password | Role |
|---|---|---|
admin |
admin123 |
admin |
developer |
developer123 |
developer |
viewer |
viewer123 |
viewer |
These are intentionally demo-only credentials. Replace the identity layer before deployment.
Run Tests
python -m pytest -q
Optional Standards-Based MCP Server
python -m pip install -r requirements-mcp.txt
python mcp_server.py
Connect an MCP client or MCP Inspector to http://localhost:8000/mcp.
Docker Compose
docker compose up --build
- API docs:
http://localhost:8000/docs - Jaeger UI:
http://localhost:16686
Documentation
See docs/ for setup, architecture, API reference, MCP notes, and security considerations.
Security Note
The REST gateway demonstrates authentication, authorization, rate limiting, auditing, and allowlisted outbound calls. The optional local MCP server is a separate SDK demonstration. A remote MCP deployment should use OAuth 2.1 resource-server protection and production identity infrastructure.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。