fastly-mcp

fastly-mcp

Enables interaction with the Fastly REST API through user-scoped authentication, supporting service management, purging, token updates, and full API coverage via a configurable request tool.

Category
访问服务器

README

fastly-mcp

A Node.js Model Context Protocol server for the Fastly REST API. It supports stdio and Streamable HTTP transports, tenant-scoped user or account identities, persistent Postgres configuration, and persistent Vault secrets.

Guarantees

  • Every Fastly request uses the caller's user-scoped Fastly-Key; no shared Fastly token is configured in the process.
  • Token material is written only to Vault at fastly/tenants/<tenant>/{users|accounts}/<principal>/http/auth/token-index/fastly.
  • Non-secret configuration uses the Postgres table fastly_config, keyed by tenant, principal type, principal ID, and key.
  • Mutations require MCP_ADMIN_AUTH_KEY through the authorizationKey tool parameter when that environment variable is set.
  • service_api_request supports every documented Fastly method and path, including APIs added after this server release.

Tools

service_query_suggestion is read-only schema discovery. Use it when intent or the correct Fastly route is unclear; it returns recommended prerequisites, safety checks, tool schemas, and examples.

service_connection_info and service_scope_info are read-only runtime and tenant/account discovery tools. Use them before an operational workflow.

service_health_check calls GET /tokens/self with a user token. fastly_list_services calls GET /service. fastly_list_versions calls GET /service/:service_id/version.

service_api_request is the full-coverage escape hatch. It accepts method, path, optional query, body, headers, and required fastlyToken. GET/HEAD are read-only; POST/PUT/PATCH/DELETE are mutations and require authorizationKey when configured.

fastly_purge_service is high-risk and calls POST /service/:service_id/purge_all. Confirm the service ID and use the admin key. fastly_update_user_token is high-risk and persists a replacement token in Vault; provide exactly one of userId or accountId, plus tenantId, fastlyToken, and the admin key.

All responses contain JSON text shaped as { "ok": true, "status": 200, "data": {} }; failures set MCP isError and return { "ok": false, "status": 4xx, "error": "..." }. Secret-looking fields are redacted by default.

Configuration

Copy .env.example to .env. Set POSTGRES_*, VAULT_ADDR, VAULT_TOKEN or Vault Agent settings, and MCP_ADMIN_AUTH_KEY. TARGET_SERVICE_BASE_URL defaults to https://api.fastly.com. Use MCP_TRANSPORT_MODE=stdio, http, or both.

Run local dependencies with docker compose up -d postgres vault, then initialize Vault according to vault-production/README.md. Start the server with npm start, or use npm run start:http for HTTP. The HTTP endpoint defaults to http://127.0.0.1:3000/mcp and requires a bearer token.

External Services Mode

Use docker-compose.external.yml when Fastly MCP connects to external Vault and Postgres services. Set POSTGRES_HOST, VAULT_ADDR, and the corresponding credentials before starting the app-only Compose service.

Inventory and tests

npm run inventory:generate fetches Fastly's API reference index, or a JSON OpenAPI document supplied by FASTLY_OPENAPI_URL, and writes artifacts/fastly-openapi-endpoints.json. The artifact is committed and CI fails when generation changes it.

npm test runs the Fastly client, MCP authorization, persistence-supporting infrastructure, HTTP, and Vault tests. GitHub Actions runs inventory generation before the test job.

MCP client registration

{
  "mcpServers": {
    "fastly-mcp": {
      "command": "npm",
      "args": ["run", "start:stdio"],
      "cwd": "/Users/{user}/Documents/GitHub/fastly-mcp"
    }
  }
}

Development notes

The Fastly API uses TLS 1.2 or later and applies separate read, write, and purge limits. Avoid concurrent writes to the same service. Never log or commit fastlyToken values. This project is released under the MIT License; see LICENSE.

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选