Quip MCP Server (TypeScript)

What is Quip MCP Server (TypeScript)?

A Model Context Protocol (MCP) server for reading spreadsheet data from Quip documents and returning it in CSV format. Implemented in TypeScript, it provides a single tool (quip_read_spreadsheet) and supports both stdio and HTTP transports. Designed for developers who want to integrate Quip spreadsheet content into AI assistants like Claude.

How to use Quip MCP Server (TypeScript)?

Install globally via npm (npm install -g @zxkane/quip-mcp-server) or from source. Set the QUIP_TOKEN environment variable with your Quip API token. Run the server with default stdio transport using npm start, or use command‑line arguments to configure storage paths, debug mode, mock mode, authentication, and HTTP/SSE options. For Claude.app, add a JSON configuration block to your MCP settings.

Key features of Quip MCP Server (TypeScript)

• Retrieve Quip spreadsheet content as CSV with metadata
• Support for selecting specific sheets by name
• Automatic truncation of large spreadsheets
• Mock mode for testing without a real Quip API token
• Multiple storage options: local filesystem and Amazon S3
• API key authentication and health check endpoint

Use cases of Quip MCP Server (TypeScript)

• Let an AI assistant pull live Quip spreadsheet data into a conversation
• Automate extraction of sheet content for reporting or analysis pipelines
• Test and develop MCP integrations without consuming Quip API quota

FAQ from Quip MCP Server (TypeScript)

How do I authenticate with Quip?

Set the QUIP_TOKEN environment variable to your Quip API token. Optionally, set QUIP_BASE_URL (defaults to https://platform.quip.com). In mock mode (--mock), no token is required.

What transport protocols are supported?

Two transport types: stdio (default for subprocess use like Claude.app) and HTTP (set PORT/HOST environment variables). The server also supports Server‑Sent Events (SSE) via --sse flag for real‑time streaming.

Can I use this server without a real Quip API token?

Yes. Use the --mock command‑line argument (or set QUIP_MOCK=true environment variable) to enable mock mode, which returns placeholder data for testing.

How can I deploy this to AWS?

The README provides two serverless deployment options: AWS Lambda + API Gateway (auto‑scaling, HTTP interface) and AWS Agent Core Runtime (SSE support, up to 8‑hour runtime). Both require a Quip token stored in AWS Secrets Manager and are deployable via provided shell scripts.

What tools does the server expose?

A single tool: quip_read_spreadsheet. It accepts a required threadId and an optional sheetName, returning CSV content and metadata (total rows, total size) as a JSON object.