Serper Search MCP Server

This repository provides a Model Context Protocol (MCP) server that integrates Google search capabilities via the Serper API, featuring an AI-powered Deep Research tool. It enables search and research functionality within MCP-enabled applications.

Features

• Google search integration through Serper API.
• Rich search response data: Knowledge Graph, organic results, "People Also Ask," related searches.
• AI-powered Deep Research: iterative research, sub-queries, information synthesis, citations, adjustable depth, quality metrics.
• Configurable search parameters: country targeting, language selection, result count, autocorrect.
• Secure API key handling, rate limiting, caching, TypeScript support, and integrated performance metrics.

Installation

Clone the repository, install dependencies using pnpm install, and build the server with pnpm run build.

Configuration

Obtain a Serper API key and create a .env file. Advanced quality metrics can be configured using USAGE_METRICS_KEY, USAGE_PROJECT_ID, METRICS_ENDPOINT, and DISABLE_METRICS. See TELEMETRY.md for details.

Integration

Add the server config to your Claude Desktop configuration file.

Usage

The server offers a search tool with parameters like query, numResults, gl, hl, autocorrect, and type. The Deep Research tool supports query, depth ("basic", "standard", "deep"), and maxSources.

Response Types

Includes Knowledge Graph, Organic Results, People Also Ask, and Related Searches.

Quality Metrics

The Deep Research tool integrates metrics for research process, performance, issue tracking, usage patterns, and result quality.

Contributing

Pull requests are welcome.

License

MIT License.

Acknowledgments

Serper API, Model Context Protocol, and PostHog.