Brewman MCP Server
Wraps the Brewman Web (V7) API to read and write Brewman data (orders, outlets, stock, config) via tools, with the API token stored securely as an environment variable.
README
brewman-mcp-server
An MCP server that wraps the Brewman Web (V7) API so Claude can read and write
Brewman data through tools — without ever seeing the raw API token. Built for Flying
Tumbler as the fast read/write layer alongside the browser-driven brewman-operator
skill (which handles production, duty, ullages, containers, distribution and reports —
the things the API can't do).
The token is held as a server environment variable. It is never in code, never in chat, never logged.
What it exposes
Tools (all calls are POST to /webapi/...; the tenant id is injected automatically):
| Tool | Type | Purpose |
|---|---|---|
brewman_get_user_context |
read | Verify token + get the tenant id. Call first. |
brewman_get_reference |
read | Any config list: product_brands, stock_locations, package_types, price_lists, pricing_categories, sales_areas/codes, credit_terms, vat_codes, gl_codes, stock_groups, units_of_measure, outlet_types/ratings/sub_types, couriers, delivery_areas, collection_times, outlet_groups, suppliers_manufacturers, outlets. |
brewman_get_stock_levels |
read | Current in-stock qty per item id. |
brewman_get_stock_incoming_outgoing |
read | Incoming/outgoing totals. |
brewman_get_price_list_lines |
read | All product prices on a price list. |
brewman_evaluate_prices |
read | Price items on a list (list price + discounts). |
brewman_get_outlets_by_filter |
read | Outlets (all or by id). |
brewman_create_outlet |
write | Create a customer. |
brewman_update_outlet |
write | Update a customer. |
brewman_get_orders_by_filter |
read | Orders (header + lines) by filter. |
brewman_build_order_lines |
read | Preview an order's lines/pricing WITHOUT saving. |
brewman_create_order |
write | Create an order (Open status). |
brewman_cancel_order |
write/destructive | Cancel an order. |
brewman_get_current_posting_items |
read | What's queued for accounts posting. |
brewman_request |
advanced | Generic passthrough to ANY of the ~90 Brewman methods. |
Full API reference (Swagger): https://brewman.premiersystems.com/swagger/index.html
Setup
1. Create a dedicated Brewman API user + token
In Brewman, create a new user just for the API (so anything it creates is attributed to "API" in the UI, not to you). Generate an API token from that user's screen and note the tenant (group) id shown there.
2. Configure environment variables
Copy .env.example → .env (local) or set them in your host's project settings:
BREWMAN_API_TOKEN— the token (secret)BREWMAN_TENANT_ID— the tenant/group id- optional:
BREWMAN_BASE_URL,BREWMAN_TENANT_FIELD,TRANSPORT,PORT
3a. Deploy to Vercel (matches the FT Meta MCP setup)
npm install
vercel # link/deploy
# In the Vercel dashboard: Project → Settings → Environment Variables:
# BREWMAN_API_TOKEN, BREWMAN_TENANT_ID
vercel --prod
MCP endpoint: https://<your-app>.vercel.app/mcp (rewritten to /api/mcp).
3b. Or run anywhere as a Node service / locally
npm install
npm run build
BREWMAN_API_TOKEN=... BREWMAN_TENANT_ID=... npm start # http on :3000 -> /mcp
4. Connect it to Claude / Cowork
Add the deployed URL as a custom MCP connector (same way the FT Meta MCP is wired):
https://<your-app>.vercel.app/mcp. Then the brewman-operator skill will use these
tools for orders/outlets/stock/config and fall back to the browser for everything else.
Verify
# health check
curl https://<your-app>.vercel.app/mcp # -> {"ok":true,...}
Then from Claude, call brewman_get_user_context — it should return your user + tenant id.
Safety
- Reads are free. Writes (
create_order,create/update_outlet,cancel_order, and any write viabrewman_request) change live data exactly like the UI — the operator confirms with Patrick before calling them. - The API is BETA; schemas may change. The dedicated tools cover the common cases; use
brewman_requestwith the Swagger schema for anything else. - Not available via API (use the browser/
brewman-operatorskill): production/batches, duty & excise returns, ullages, container tracking, distribution journeys, BI reports.
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。