Perplexity MCP Server

Name: Perplexity MCP Server
Author: wysh3

@wysh3

代码仓库

关于 Perplexity MCP Server

MCP web search using perplexity without any API KEYS

#perplexity #mcp-server

基本信息

分类

搜索

许可证

GPL-3.0

运行时

node

传输方式

stdio

发布者

wysh3

配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "perplexity-mcp-server": {
      "command": "bun",
      "args": [
        "run",
        "build"
      ]
    }
  }
}

工具

未检测到工具

工具是从 README 中自动提取的。维护者可以在 ## Tools 标题下列出工具,即可填充这部分内容。

概览

What is Perplexity MCP Server?

Perplexity MCP Server is a minimalist research server implementing the Model Context Protocol (MCP) to deliver AI-powered research capabilities through Perplexity's web interface. It uses browser automation to perform searches, extract content, and maintain persistent conversations without requiring an API key. It is designed for developers and researchers who need intelligent web research, documentation retrieval, and code analysis.

How to use Perplexity MCP Server?

Install the Bun runtime and Node.js 18+, clone the repository, run bun install and bun run build. Configure your MCP client with the absolute path to the built main.js file and an optional timeout of 300 seconds. Optionally, for Pro account benefits, run bun run login and log in via email. Initiate tasks through your MCP client using natural language commands such as "Use perplexity to research quantum computing advancements" or "Ask perplexity-server for React 18 documentation."

Key features of Perplexity MCP Server

Intelligent web search without API limits
Persistent conversations via SQLite chat storage
Content extraction with GitHub repository parsing
Documentation retrieval, API discovery, and code analysis
Keyless operation through browser automation
Optional Perplexity Pro account support for better models

Use cases of Perplexity MCP Server

Research emerging technologies like quantum computing
Retrieve up‑to‑date technical documentation (e.g., React 18)
Discover relevant APIs for a development project
Analyze code snippets for deprecated patterns
Maintain context across multiple research queries

FAQ from Perplexity MCP Server

What does Perplexity MCP Server do differently from traditional APIs?

It requires no authentication or API key – it works through browser automation. It is free, processes data locally, includes native GitHub integration, and stores conversation history in SQLite.

What are the runtime requirements?

Bun runtime and Node.js 18+ (for TypeScript compilation) are required.

How does data privacy work?

All processing is local; conversation history is stored in a local SQLite database. No data is sent to remote servers except through Perplexity’s web interface.

How do I use my Perplexity Pro subscription?

Run bun run login, log in via email (recommended for best compatibility) in the browser window that opens, then close the browser. The session is saved and will be used automatically.

Are there any known limitations?

Stability depends on Perplexity’s website consistency. GitHub paths must use full repository URLs. Link recursion depth can be adjusted in the source configuration. The project is licensed under GNU GPL v3.0.