sellercloud-mcp
The first Model Context Protocol server for SellerCloud. Plug Claude into your catalog, inventory, orders, and channel listings — read-only, in five minutes.
README
sellercloud-mcp
The first Model Context Protocol server for SellerCloud. Plug Claude into your catalog, inventory, orders, and channel listings — read-only, in five minutes.
Why this exists
SellerCloud has no public SDK. Their REST API is well-documented but unbranded — every team that automates against it ends up writing the same auth-and-pagination glue from scratch.
If you use Claude (or any MCP-aware AI assistant) to operate ecommerce day-to-day, that gap is the difference between "summarize today's orders" working out of the box and "summarize today's orders" requiring a custom integration.
sellercloud-mcp closes that gap. It's a tiny, well-tested, MIT-licensed MCP server that exposes seven read-only SellerCloud endpoints to any MCP client. Built from years of running ecommerce automation at scale.
What you can do with it
Wire this server into Claude Code, Claude Desktop, or any MCP host, then ask things like:
- "Search for any SKU containing
WIDGETand show me the inventory levels." - "How many orders did we ship yesterday across all marketplaces? Group by channel."
- "Pull order 100001 and tell me which line items shipped."
- "List the channels configured for company 9001 and show which ones are active."
- "For SKU
ACME-001, compare the price across every channel listing."
Claude reads your catalog directly. No copy-paste, no spreadsheets, no custom pipelines.
Tools (v0.1, all read-only)
| Tool | What it does |
|---|---|
sellercloud_search_products |
Free-text search across catalog (name, SKU, attributes). |
sellercloud_get_product |
Fetch one product by exact SKU. |
sellercloud_search_orders |
List orders in a date window, optionally scoped by company. |
sellercloud_get_order |
Fetch one order by ID, including line items. |
sellercloud_get_inventory |
Current on-hand / reserved / on-order qty for one SKU. |
sellercloud_list_channels |
List configured marketplace/channel feeds. |
sellercloud_get_channel_listing |
Per-channel listing detail for one SKU. |
Write endpoints (create order, update inventory, push channel changes) are intentionally not in v0.1. They are planned for v0.2 once read-only ergonomics settle.
Install
pip install sellercloud-mcp
v0.1 ships from this repository. PyPI publication is pending — for now, install with
pip install git+https://github.com/alveyautomation/sellercloud-mcpor clone and runpip install -e .locally.
Configure credentials
The server reads everything from environment variables. Copy .env.example to .env and fill in your tenant:
SELLERCLOUD_API_URL=https://your-team.api.sellercloud.com/rest/
SELLERCLOUD_USERNAME=your-username
SELLERCLOUD_PASSWORD=your-password
SELLERCLOUD_DEFAULT_COMPANY_ID= # optional fallback
SELLERCLOUD_HTTP_TIMEOUT=60 # optional, seconds
SELLERCLOUD_MAX_RETRIES=3 # optional
Use a read-only SellerCloud account. v0.1 only calls
GETendpoints, but defense in depth means you should hand the server a dedicated user that cannot modify anything. When v0.2 lands with write tools, opt-in by upgrading the credential — never the other way around.
Wire into Claude Code
Add to ~/.claude/claude_code_config.json (or your project's MCP config):
{
"mcpServers": {
"sellercloud": {
"command": "sellercloud-mcp",
"env": {
"SELLERCLOUD_API_URL": "https://your-team.api.sellercloud.com/rest/",
"SELLERCLOUD_USERNAME": "your-username",
"SELLERCLOUD_PASSWORD": "your-password",
"SELLERCLOUD_DEFAULT_COMPANY_ID": "9001"
}
}
}
}
Restart Claude Code. The seven sellercloud_* tools will appear in any new session.
Wire into Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows) and add the same mcpServers block as above. Restart the desktop app.
Tool reference
Every tool returns a JSON envelope:
{ "ok": true, "data": { ... } }
{ "ok": false, "error": "human-readable message" }
sellercloud_search_products
sellercloud_search_products(
query: str, # required
company_id: int | None = None, # falls back to default if unset
page: int = 1,
page_size: int = 50, # capped at 50 by SellerCloud
)
Example response:
{
"ok": true,
"data": {
"items": [
{ "ID": "ACME-WIDGET-001", "ProductName": "Acme Widget, Standard", "Price": 29.99 }
],
"total": 1,
"page": 1,
"page_size": 50
}
}
sellercloud_get_product
sellercloud_get_product(sku: str, company_id: int | None = None)
Returns the catalog record, or data: null if the SKU is not in the company's catalog.
sellercloud_search_orders
sellercloud_search_orders(
date_from: str, # ISO date "YYYY-MM-DD"
date_to: str, # ISO date "YYYY-MM-DD"
company_id: int | None = None,
query: str | None = None,
limit: int = 200, # max 1000
)
Pagination is handled transparently — SellerCloud caps page size at 50, but the tool collects pages up to limit. The response includes limit_reached: true when there were more orders than limit allowed.
sellercloud_get_order
sellercloud_get_order(order_id: int)
Returns the full order record (with Items[]), or data: null for a 404.
sellercloud_get_inventory
sellercloud_get_inventory(sku: str, company_id: int | None = None)
The returned record includes:
InventoryAvailableQty— what the API considers sellable right nowPhysicalQty— on-handReservedQty— held for open ordersOnOrder— incoming PO qty
Use InventoryAvailableQty as the canonical "qty I can sell" number.
sellercloud_list_channels
sellercloud_list_channels(company_id: int | None = None)
Returns the list of configured channel feeds for the company. Each record includes ChannelID, Name, and Active.
sellercloud_get_channel_listing
sellercloud_get_channel_listing(channel_id: int, sku: str)
Per-channel listing detail. Useful for spot-checking prices across marketplaces.
Local development
git clone https://github.com/alveyautomation/sellercloud-mcp
cd sellercloud-mcp
python -m venv .venv && source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install -e ".[dev]"
pytest # 44 tests, ~4s
Pre-commit hooks (gitleaks, ruff, formatter, tenant-fingerprint scrubber):
pip install pre-commit
pre-commit install
Integration tests against a real SellerCloud sandbox account are gated behind SELLERCLOUD_INTEGRATION_TESTS=1. They are not required for normal contribution.
Troubleshooting
Failed to obtain SellerCloud token — username/password rejected. Most common cause: the account has 2FA enabled or is locked out. SellerCloud's POST /api/token endpoint expects a non-2FA service account.
Missing required environment variables — the server tried to start before its .env was loaded. Either export the vars in the parent shell, or ensure your MCP host config includes them in the env block.
Empty results despite known data — confirm the company_id is correct. SellerCloud returns only the authenticated user's default company unless you pass companyID explicitly.
Pagination feels slow — page size is capped at 50 by SellerCloud, not by us. For large date windows, expect multiple round-trips.
Contributing
Issues and pull requests welcome. Please:
- Run
pytestbefore opening a PR (pip install -e ".[dev]"). - Run
pre-commit run --all-files. - Keep additions to v0.1 scope read-only. Write endpoints land in v0.2.
- Synthetic data only in tests — no real SKUs, customer names, or order numbers.
License
MIT — see LICENSE.
Disclaimer
sellercloud-mcp is an unofficial, third-party integration. It is not endorsed by, affiliated with, or supported by SellerCloud, Inc. "SellerCloud" is a trademark of SellerCloud, Inc. Use at your own risk; verify behavior against your tenant before depending on it for production decisions.
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。