Model Context Protocol

Integrate Data Natively Into Chat

Overview

Birkini supports Anthropic’s Model Context Protocol (MCP), enabling seamless integration of structured data directly into chat-based interfaces.

This allows teams to enrich their existing chat experiences with verified datasets from the marketplace as well as proprietary organizational data.

More information on this feature is available in the release notes.

Step 1: Create your account Go to the Birkini platform and sign up using email or SSO.

Step 2: Choose your storage provider Select your preferred provider and region, then provision your data warehouse.

Step 3: Subscribe to a dataset (optional) In the sidebar, open the Marketplace and subscribe to any available dataset.

Step 4: Install MCP for Claude Desktop Follow the latest instructions in the official repository to install MCP.

Step 5: Restart and start asking questions Launch Claude Desktop with MCP enabled and begin querying data contextually.

Setup in Cursor

Step 1: Complete the steps above, and make sure Python 3.11 or higher and uv are installed. On macOS, use brew install uv.

Step 2: Open Cursor Settings.

Step 3: Select Add a new MCP server. Give it the name Birkini and keep the type set to command.

Step 4: In the command field, paste the following line:

env db_session_key=YOUR_DB_SESSION_KEY uvx birkini-mcp

Step 5: You can now interact with Birkini through Cursor chat using your configured MCP server.