perplexityai/modelcontextprotocol

What it solves

This project provides an official Model Context Protocol (MCP) server implementation for the Perplexity API Platform. It allows AI assistants (like Cursor, VS Code, or Claude Desktop) to access real-time web search, advanced reasoning, and deep research capabilities directly within their interface by connecting to Perplexity's infrastructure.

How it works

The server acts as a bridge between an MCP-compatible client and the Perplexity API. It can be deployed as a remote server (hosted by Perplexity) for instant connectivity or as a local server (via npx or Docker) for more control. It exposes four primary tools to the AI assistant:

• perplexity_search: Performs direct web searches for current information with metadata.
• perplexity_ask: A conversational AI tool for quick questions using a "fast" preset.
• perplexity_research: A deep research tool for comprehensive analysis and reports using a "high" preset.
• perplexity_reason: An analytical tool for complex problem-solving using a "medium" preset.

Who it’s for

Developers and users of AI assistants and IDEs (such as Cursor, VS Code, Claude Desktop, and Windsurf) who want to integrate real-time web search and research capabilities into their AI-powered workflows.

Highlights

• Multiple Deployment Options: Supports remote HTTP, local stdio, and self-hosted HTTP mode via Docker or Node.js.
• Enterprise Ready: Includes built-in support for corporate proxies (PERPLEXITY_PROXY).
• Flexible Integration: Can be used as a standalone server or imported as a library into other Node.js processes.
• Detailed Tooling: Provides specialized tools for different intensities of search and reasoning (fast, medium, high).