gas-mcp
Turns a GAS service into an MCP server, exposing tools for session creation, resource reading, searching, and executing capabilities.
README
gas-mcp
gas-mcp is the generic MCP transport for a
gas_protocol.service.GasService. It turns one GAS service into an MCP server
without knowing which domain, database, policy implementation, or application
created that service.
What it provides
install_gas_mcp(server, service) registers five structured-output MCP tools:
create_session— create an isolated session and return its initial view.get— read a resource URI and its contextual commands and links.search— search a resource collection.act— execute an advertised capability with a revision and idempotency key.why_not— diagnose an unavailable command without returning an executable command.
The installer adds read-only or mutation tool annotations and translates
GasError failures into MCP errors. Domain-specific resource names and input
schemas stay in the GasBackend and its composition root.
The package provides transport only. The GAS request/response contract and
capability semantics remain defined by gas-protocol; authorization and
domain policy remain with the backend or application that supplies the
service.
Install
pip install "gas-mcp[mcp]"
The mcp extra supplies the MCP SDK needed to import and run the transport.
Example
from mcp.server import MCPServer
from gas_mcp import install_gas_mcp
from gas_protocol import GasService, InMemoryGasBackend
server = MCPServer(name="example")
install_gas_mcp(server, GasService(InMemoryGasBackend()))
For a real application, construct the GasService over its own backend and
keep domain-specific MCP resources in that application's composition root.
Contributing
uv build
uv run --extra mcp --with pytest pytest -q
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。
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。
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。
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。
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。
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。
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。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。