Bring! Shopping MCP Server
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Bring! Shopping MCP Serveradd milk and eggs to my grocery list"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP Server for Bring! Shopping

This project implements a local Model Context Protocol (MCP) server in TypeScript that exposes the functionalities of the Bring! shopping list API. It enables applications like Claude Desktop to interact with your Bring! shopping lists using standardized MCP tools.
The server integrates the bring-shopping npm package for Bring! API access and uses the MCP TypeScript SDK v2 server package to provide an MCP-compliant interface.
Disclaimer:
This is a personal project. I am not affiliated with Bring! Labs AG in any way.
This project uses an unofficial Bring! API, which may change or be blocked at any time.
This could cause the MCP server to stop functioning without prior notice.
๐งฉ Recommended Claude Desktop Configuration
To use this server in Claude Desktop via npx, insert the following into your claude_desktop_config.json file:
{
"mcpServers": {
"bring-mcp": {
"command": "npx",
"args": ["-y", "bring-mcp@latest"],
"env": {
"BRING_EMAIL": "your_bring_email@example.com",
"BRING_PASSWORD": "YOUR_BRING_PASSWORD_HERE"
}
}
}
}This is the recommended and most portable configuration. It ensures you always use the latest version published to npm without needing local installation.
Related MCP server: lista-compras-mcp
๐ฌ Using with ChatGPT via OpenAI Secure MCP Tunnel
OpenAI Secure MCP Tunnel lets ChatGPT reach this STDIO server without exposing it through a public endpoint. The tunnel client itself is provided as a downloadable binary, while bring-mcp can still be launched through npx without a local checkout or installation.
Prerequisites
Node.js 22 or newer, including
npm/npxChatGPT developer-mode access
Tunnels Read + Usepermissions in the relevant OpenAI Platform organizationA tunnel ID and runtime API key from OpenAI Platform tunnel settings
The
tunnel-clientbinary, available from the download link in the tunnel settings
When creating the tunnel, associate it with the ChatGPT workspace in which you want to use Bring. Otherwise, it will not appear when creating the ChatGPT connection.
Configure and run the tunnel
Export the tunnel API key and your Bring! credentials in the shell that will run tunnel-client:
export CONTROL_PLANE_API_KEY="sk-..."
export BRING_EMAIL="your_bring_email@example.com"
export BRING_PASSWORD="YOUR_BRING_PASSWORD_HERE"Create a tunnel profile that starts the latest published bring-mcp package through npx:
tunnel-client init \
--sample sample_mcp_stdio_local \
--profile bring-mcp \
--tunnel-id tunnel_0123456789abcdef0123456789abcdef \
--mcp-command "npx -y bring-mcp@latest"Validate the configuration, then start the tunnel:
tunnel-client doctor --profile bring-mcp --explain
tunnel-client run --profile bring-mcpKeep this process running while using the integration. The environment variables above are inherited by the npx-started MCP server.
Connect the tunnel in ChatGPT
In ChatGPT, open Settings โ Security and login and enable Developer mode.
Open ChatGPT Plugins and select the plus button to create a developer-mode app.
Enter a name and description, choose Tunnel under Connection, and select the tunnel created above (or enter its
tunnel_id).Create the connection and review the 16 discovered Bring! tools.
Start a new chat, enable the Bring connection from the tools menu, and try a request such as โShow my Bring shopping lists.โ
If ChatGPT cannot discover the server, make sure tunnel-client run is still active, rerun the doctor command, and verify that the tunnel is associated with the correct ChatGPT workspace. Secure MCP Tunnel is intended for private connections and developer-mode testing; publishing a public ChatGPT plugin requires a stable public HTTPS MCP endpoint.
๐ Features
Automatic Authentication: No manual login required - authentication happens automatically on first API call
Exposes Bring! API functions as MCP tools:
๐งพ Load shopping lists
๐ Get and modify items (add, remove, move)
๐ฆ Batch operations (save multiple items, delete multiple items)
๐ผ Save/remove item images
๐ฅ Manage list users
๐ฏ Get default shopping list UUID
๐ Load translations & catalog
๐จ Retrieve pending invitations
Communicates via STDIO (for use with Claude Desktop or MCP Inspector)
Supports MCP protocol revision
2026-07-28while continuing to serve legacy 2025 clientsPublishes tool titles, annotations, concrete input/output schemas, and machine-readable structured results
Marks tool failures with
isError: trueso clients can distinguish them from successful callsSupports Bring! credentials via
.envfile or injected environment variables
Available Tools
loadLists: Load all shopping lists from Bring!getItems: Get all items from a specific shopping listgetItemsDetails: Get details for items in a listsaveItem: Save an item to a shopping list with optional specificationsaveItemBatch: Save multiple items to a shopping list in one operationremoveItem: Remove an item from a specific shopping listmoveToRecentList: Move an item to the recently used items listdeleteMultipleItemsFromList: Delete multiple items from a list by their namessaveItemImage: Save an item image from base64-encoded image data (maximum decoded size: 5 MiB)removeItemImage: Remove an image from an itemgetAllUsersFromList: Get all users associated with a shopping listgetUserSettings: Get settings for the authenticated usergetDefaultList: Get the UUID of the default shopping list (use when user doesn't specify a list)loadTranslations: Load translations for the Bring! interfaceloadCatalog: Load the Bring! item cataloggetPendingInvitations: Get pending invitations to join shopping lists
โ๏ธ Setup and Installation
Clone the repo (or obtain the files)
Navigate into the project directory:
cd path/to/bring-mcpInstall dependencies:
npm installCreate
.envfile (if not injecting ENV directly):BRING_EMAIL=your_email@example.com BRING_PASSWORD=your_passwordBuild the project:
npm run buildMake script executable (optional on Unix):
chmod +x build/src/index.js
๐ Running the Server
Launch the MCP server with:
node build/src/index.jsIf successful, you'll see: MCP server for Bring! API v<version> is running on STDIO (on stderr).
๐งช Testing with MCP Inspector
Ensure
npm run buildhas been executed.Ensure
.envwith valid credentials exists.Run Inspector:
npx @modelcontextprotocol/inspector node /ABS/PATH/bring-mcp/build/src/index.js
๐งฉ Claude Desktop Integration (Manual Local Setup)
Alternatively, if you prefer a locally built and installed version:
{
"mcpServers": {
"mcp-bring": {
"command": "node",
"args": ["/ABSOLUTE/PATH/TO/bring-mcp/build/src/index.js"],
"env": {
"BRING_EMAIL": "your_bring_email@example.com",
"BRING_PASSWORD": "YOUR_BRING_PASSWORD_HERE"
}
}
}
}๐ง Development
Testing
Run tests with:
npm run testThis command runs formatting, linting, and Jest tests with coverage reporting.
For CI testing:
npm run test:ciBuilding
Build the project:
npm run buildKey Dependencies and Tools
@modelcontextprotocol/server: MCP SDK v2 server and STDIO protocol dispatcher@modelcontextprotocol/client: Development-only client used for end-to-end protocol compatibility tests@modelcontextprotocol/inspector: Run on demand withnpxfor testing and debugging MCP serversbring-shopping: Node.js wrapper for the Bring! APIzod: For schema definition and validationdotenv: For managing environment variables
โ Final Notes
๐ Avoid committing your
.envfile.โป๏ธ
MAILandPWremain supported as deprecated aliases for existing installations.๐งผ Keep credentials out of version control.
๐ MCP Inspector is invaluable for debugging.
๐ Authentication is handled automatically - no manual login required.
๐ฆ Use batch operations for efficiency when working with multiple items.
Happy coding with MCP and Bring! ๐
This server cannot be deployed
Maintenance
Related MCP Connectors
Manage Superlist tasks and lists in plain language from any MCP-compatible AI agent.
Turn any shopping list into a ready-to-checkout grocery cart across 26 European supermarkets.
Persistent shopping memory for AI assistants. The AI reasons; Kaufliste stores and executes.
Household-aware cooking brain: pantry, meal suggestions, dietary safety, recipes, shopping lists.
Related MCP Servers
- FlicenseNot gradedqualityCmaintenanceMCP server for the Bring! shopping list API, enabling management of shopping lists via natural language.-
- FlicenseAqualityDmaintenanceEnables AI assistants to manage shopping lists and items (create, edit, delete, mark as purchased) via integration with a backend API.8-
- AlicenseNot gradedqualityDmaintenanceEnables AI assistants to read and manage your Bring! shopping lists through natural language commands like adding items, marking purchases, and retrieving list contents.MIT
- AlicenseNot gradedqualityAmaintenanceEnables verified Bring! shopping-list operations such as listing items, searching the catalog, classifying products, adding/moving/checking/removing items, and managing product photos through MCP tools.58 npm1MIT