squarespace-mcp
Enables LLM agents to interact with Squarespace Commerce APIs (products, orders, inventory, transactions) using natural language.
README
squarespace-mcp
An MCP server that exposes the Squarespace Commerce API as tools an LLM agent can call — read products, inventory, orders, transactions, and customer profiles, and (with care) adjust inventory and fulfill orders.
- Stack: Python + FastMCP 2.x/3.x
- Transport: stdio
- API: Squarespace Commerce APIs (
https://api.squarespace.com), Bearer-key auth
Setup
# with uv (recommended)
uv venv && source .venv/bin/activate
uv pip install -e ".[dev]"
# or plain pip
python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"
cp .env.example .env # then fill in SQUARESPACE_API_KEY
Get an API key from Squarespace → Settings → Developer API keys. The key is
read from the environment (or a local .env) lazily — the server starts and
registers its tools even with no key set; a clear error is raised only when a
tool is actually invoked without one.
Environment variables
| Variable | Required | Default | Purpose |
|---|---|---|---|
SQUARESPACE_API_KEY |
yes (at call time) | — | Bearer API key |
SQUARESPACE_API_VERSION |
no | 1.0 |
Version for inventory / orders / transactions / profiles |
SQUARESPACE_PRODUCTS_API_VERSION |
no | v2 |
Products is versioned independently — v2 is current (legacy 1.0/1.1 not for new integrations) |
SQUARESPACE_USER_AGENT |
no | squarespace-mcp/0.1 … |
Squarespace requires a descriptive User-Agent |
Run
squarespace-mcp # runs the MCP server over stdio
The process reads/writes MCP frames on stdin/stdout and blocks waiting for a client — that is expected. It is meant to be launched by an MCP client, not run interactively.
Register with Claude Code
claude mcp add squarespace \
--env SQUARESPACE_API_KEY=your_key_here \
-- /absolute/path/to/squarespace-mcp/.venv/bin/squarespace-mcp
(Or point the command at uv run squarespace-mcp / your installed console
script.) After registering, the tools below appear to the agent.
Tools
Read (safe)
| Tool | Purpose |
|---|---|
list_products(cursor?, modified_after?) |
List catalog products, 50/page |
get_product(product_id) |
Retrieve one product |
list_inventory(cursor?) |
List stock levels, 50/page |
get_inventory(variant_ids) |
Stock for specific variants (≤50 ids) |
list_orders(cursor?, modified_after?, fulfillment_status?) |
List orders (filter PENDING/FULFILLED/CANCELED) |
get_order(order_id) |
Retrieve one order |
list_transactions(cursor?, modified_after?, modified_before?) |
List financial transactions |
list_profiles(cursor?) |
List customer profiles |
get_profile(profile_id) |
Retrieve one profile |
Pagination: list tools return the raw Squarespace page including a
pagination object; pass its nextPageCursor back as cursor to page forward.
Write (⚠️ mutates the live store)
| Tool | Purpose |
|---|---|
adjust_inventory(variant_id, quantity) |
Set a variant's stock to an exact quantity (sent with an idempotency key) |
fulfill_order(order_id, shipments, should_send_notification=true) |
Mark an order fulfilled with shipment details; optionally emails the customer |
Destructive operations (product delete, etc.) are intentionally not implemented.
Development
pytest -q # all HTTP is mocked with respx — no network calls
Security
- Real credentials live in
.env(gitignored). Only.env.exampleis committed. - Never hardcode the API key; it is read from the environment at runtime.
- The client raises on any non-2xx response (status + body) — no silent fallbacks.
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。