Connect Sipseek MCP

Connect Sipseek to ChatGPT, Claude, Cursor, or any MCP-compatible app and start discovering drinks.

Endpoint: https://sipseek.com/mcp

For a quick overview, visit Sipseek for AI. The full AI reference is available in llms.txt.

1. Add the server in your client

ChatGPT web:

Open Settings → Security and login and turn on Developer mode.
Open ChatGPT Plugins and select the plus button.
Create a connection named Sipseek with the description Search the Sipseek beverage catalog.
Under Connection, enter https://sipseek.com/mcp and choose No Authentication.
Create the connection and confirm that ChatGPT discovers the Sipseek tools.
In a new chat, open the plus menu, choose Developer mode, and select Sipseek.

No plugin-directory submission is required. Developer mode availability depends on your ChatGPT plan and workspace policy. See the official OpenAI setup guide.

Cursor — one click: Add Sipseek to Cursor

Or add it to your MCP config by hand:

{
        "mcpServers": {
          "sipseek": {
            "url": "https://sipseek.com/mcp"
          }
        }
      }

Claude Code:

claude mcp add --transport http sipseek https://sipseek.com/mcp

Claude.ai — add a custom connector pointing at https://sipseek.com/mcp.

2. Start discovering

Once connected, ask for a drink in natural language or provide factual requirements such as a category, flavor, ingredient, caffeine status, or explicit attribute. Sipseek searches all 1,000+ active beverage catalog records by default, including drinks, powders, mixes, ingredients, accessories, and beverage-focused retailers.

Use search_beverage_brands for ordinary searches and broad catalog browsing.
Use get_beverage_brand with a search result's stable ID to open a complete profile.
Use list_catalog_facets only when you need to explore the catalog's current factual filter values.
Share each brand’s sipseek_url in the answer.

Search uses one active beverage catalog scope. It includes every active record while preserving each record's precise category, subtype, format, and internal classification. Legacy entity-type restrictions are ignored, record_count is the authoritative catalog size, and archived records remain outside discovery. Natural-language text searches factual catalog fields; supported structured filters are enforced exactly, and an empty query browses the catalog with stable cursor pagination.

For subjective or occasion-based requests, the host AI should translate the request into concrete beverage terms before searching. context may help break ties, but it cannot create a factual match. If no factual requirement can be derived, omit the query to browse instead. Retired use-case and nutrition-threshold inputs are not supported and fail closed when sent by a cached client.

For nutrition, ingredients, pricing, or availability, open the product details and source links. Check the current product label or brand site for the latest information.

Sipseek is ready to use as soon as you connect. For custom or higher-volume integrations, see the Data License or contact hello@sipseek.com.