home / mcp / brewteco mcp server
Provides access to Brewteco data sources and actions via MCP tools for Claude Desktop.
Configuration
View docs{
"mcpServers": {
"apiccelli-brew_mcp": {
"command": "node",
"args": [
"/absolute/path/to/brewteco-mcp-server/dist/index.js"
],
"env": {
"BREWTECO_API_URL": "http://localhost:3700/api/v1"
}
}
}
}You can connect Brewteco data APIs to Claude Desktop using the MCP (Model Context Protocol). This server exposes five core Brewteco APIs as tools, enabling natural language queries to retrieve sales, products, team performance, customer segments, and individual profiles, and to drive automated actions from Claude.
You interact with the Brewteco MCP Server through an MCP client like Claude Desktop. After you configure the MCP connection, you can ask Claude to fetch data, compare metrics, filter customers, or view detailed profiles. Claude will call the corresponding tools in the background, pass the required parameters, and present consolidated results back to you.
Typical usage patterns include obtaining a sales summary for a specific store and period, ranking products by performance, evaluating seller performance, filtering customers by behavior, and retrieving a complete customer profile. You can combine multiple tools in a single request to build dashboards, run analyses, or drive targeted marketing actions.
Prerequisites: Node.js >= 18.0.0 and the Brewteco API running on port 3700.
Install dependencies for the MCP server.
npm installBuild the MCP server to compile TypeScript into JavaScript and generate the dist directory. The executable will be dist/index.js.
npm run buildStart the MCP server after building.
npm startOptionally, create a .env file to point the MCP server at Brewteco API endpoints.
BREWTECO_API_URL=http://localhost:3700/api/v1Configure Claude Desktop to load Brewteco MCP as a local stdio server. You must provide the absolute path to the compiled entry point and the API URL.
{
"mcpServers": {
"brewteco": {
"command": "node",
"args": ["/absolute/path/to/brewteco-mcp-server/dist/index.js"],
"env": {
"BREWTECO_API_URL": "http://localhost:3700/api/v1"
}
}
}
}The MCP server is stateless and does not persist client data. All requests are validated, timeouts are enforced, and transient retries are applied for reliability.
These examples show practical tool usage with Claude after the MCP server is configured.
Dashboard example: Claude summarizes December performance for a store by calling multiple tools and compiling a manager-friendly report.
Project structure includes source files for the MCP server, compilation targets in dist, and scripts for development, building, starting, and testing with MCP Inspector.
Fetches general sales data for a store within a specified period, with optional store filtering.
Compares sales across all stores for a given date range.
Provides a ranking of products by sales volume with optional filters such as period, category, store, and limit.
Lists all product categories.
Generates a sales performance ranking for the team with mix analysis, filtered by period and optional store.
Delivers detailed sales data by category for a single employee over a date range.
Filters and searches customers by behavior, spending, visit history, and store filters.
Provides the complete profile for a specific customer, identified by ID, CPF, or contact information.