Research Assistant

🚀 Perplexity MCP Server

An intelligent research assistant powered by Perplexity's dedicated AI model. It features an automatic query complexity detection function, which can route requests to the most suitable model for optimal results. Unlike the official server, it has search capabilities for each task.

🚀 Quick Start

The Perplexity MCP Server is an intelligent research assistant that uses Perplexity's dedicated AI model. It can automatically detect query complexity and route requests to the most appropriate model, offering search capabilities for each task.

✨ Features

• Automatic Query Complexity Detection: Automatically analyzes query complexity and routes requests to the most suitable model.
• Multiple Search Tools: Offers different tools for various types of queries, including simple searches, complex reasoning, and in - depth research.
• Flexible Configuration: Can be configured locally or using NPX.

💻 Usage Examples

Basic Usage

Here are some examples of using different tools provided by the server:

1. Search (Sonar Pro)

A quick search for simple queries and basic information lookup. It's best for straightforward questions that require concise and direct answers.

const result = await use_mcp_tool({
  server_name: "perplexity",
  tool_name: "search",
  arguments: {
    query: "法国的首都是哪里？",
    force_model: false // Optional: Force using this model even if the query seems complex
  }
});

2. Reasoning (Sonar Reasoning Pro)

Handles complex, multi - step tasks that require detailed analysis. Ideal for explanations, comparisons, and problem - solving.

const result = await use_mcp_tool({
  server_name: "perplexity",
  tool_name: "reason",
  arguments: {
    query: "比较并对比 REST 和 GraphQL API，说明它们的优缺点",
    force_model: false // Optional: Force using this model even if the query seems simple
  }
});

3. Deep Research (Sonar Deep Research)

Conducts comprehensive research and generates detailed reports. Suitable for in - depth analysis of complex topics.

const result = await use_mcp_tool({
  server_name: "perplexity",
  tool_name: "deep_research",
  arguments: {
    query: "量子计算对密码学的影响",
    focus_areas: [
      "后量子加密算法",
      "量子威胁的时间表",
      "实际缓解策略"
    ],
    force_model: false // Optional: Force using this model even if the query seems simple
  }
});

Advanced Usage

The server automatically analyzes query complexity to route requests to the most suitable model:

1. Simple Queries → Sonar Pro

   • Basic information lookup
   • Straightforward questions
   • Quick fact confirmation

2. Complex Queries → Sonar Reasoning Pro

   • How/why questions
   • Comparisons
   • Step - by - step explanations
   • Problem - solving tasks

3. Research Queries → Sonar Deep Research

   • In - depth analysis
   • Comprehensive research
   • Detailed investigations
   • Multi - faceted topics

You can override the automatic selection by setting force_model: true in the parameters of any tool.

📦 Installation

Prerequisites

• Node.js (from nodejs.org)
• Perplexity API key (from perplexity.ai/settings/api)
• Locally clone the code repository

Configuration

Add the following to your MCP settings file (location varies by platform):

{
  "mcpServers": {
    "perplexity": {
      "command": "node",
      "args": ["/path/to/perplexity-server/build/index.js"],
      "env": {
        "PERPLEXITY_API_KEY": "YOUR_API_KEY_HERE"
      },
      "disabled": false,
      "autoApprove": []
    }
  }
}

Or use NPX without local installation (recommended for macOS):

{
  "mcpServers": {
    "perplexity": {
      "command": "npx",
      "args": [
        "-y",
        "perplexity-mcp"
      ],
      "env": {
        "PERPLEXITY_API_KEY": "your_api_key"
      }
    }
  }
}

⚠️ Important Note

Due to implementation differences, the Reason tool may time out when using cline compared to other tools (e.g., cursor). However, the Reason tool can compensate for this shortcoming.