🔍 Perplexity Web-Search MCP

Supercharge your AI assistant with real-time web search and academic research capabilities

An MCP (Model Context Protocol) server that integrates Perplexity's powerful Sonar API, giving your AI assistant access to current information from the web and scholarly sources.

✨ Features

• 🌐 Real-time web search - Get current information from across the internet

• 🎓 Academic search - Access peer-reviewed papers and scholarly sources

• 📍 Location-aware results - Filter by city and country

• ⏰ Time-filtered search - Get recent results (day, week, etc.)

• 🧠 Multiple AI models - Choose from various Perplexity Sonar models

• 🔌 Universal compatibility - Works with any MCP-compatible system

Related MCP server: Perplexity API Platform MCP Server

🤖 Compatible AI Systems

This MCP server works seamlessly with:

• Claude

• Amazon Q

• Google Gemini

• OpenAI Codex

• Any MCP-compatible AI assistant

🚀 Exemplary Amazon Q Developer Setup

Here's how to properly integrate this MCP server with Amazon Q Developer:

1. Locate your Amazon Q agent configuration

~/.aws/amazonq/cli-agents/dev.json

2. Add the sonar MCP server to your configuration

{
  "$schema": "https://raw.githubusercontent.com/aws/amazon-q-developer-cli/refs/heads/main/schemas/agent-v1.json",
  "name": "dev",
  "description": "",
  "mcpServers": {
    "sonar": {
      "command": "/Users/your-username/SoftwareProjects/perplexity_sonar_mcp/.venv/bin/python",
      "args": ["/Users/your-username/SoftwareProjects/perplexity_sonar_mcp/main.py"],
      "cwd": "/Users/your-username/SoftwareProjects/perplexity_sonar_mcp",
      "env": {
        "PERPLEXITY_API_KEY": "pplx-your-api-key-here"
      }
    }
  },
  "tools": [
    "fs_read",
    "fs_write",
    "execute_bash",
    "use_aws",
    "@sonar"
  ]
}

3. Usage Example

[dev] > Web search: "Is AI a bubble?"

🛠️  Using tool: web_search from mcp server sonar
 ⋮
 ● Running web_search with the param:
 ⋮  {
 ⋮    "name": "web_search",
 ⋮    "arguments": {
 ⋮      "query": "Is AI a bubble?"
 ⋮    }
 ⋮  }

Allow this action? Use 't' to trust (always allow) this tool for the session. [y/n/t]: y

 ⋮
 ● Completed in 9.990s

> Based on current market analysis, AI is widely considered to be in a bubble phase...

Installation

With uv (recommended)

uv sync

With pip

pip install -e .

MCP Configuration

Add to your MCP client configuration:

With uv:

• Transport Type: STDIO

• Command: uv

• Arguments: run main.py

With Python:

• Transport Type: STDIO

• Command: python

• Arguments: main.py

Environment Setup

Set your Perplexity API key:

export PERPLEXITY_API_KEY="your-api-key-here"

Usage

With uv

uv run python main.py

With Python

python main.py

Requirements

• Python >=3.10

• PERPLEXITY_API_KEY environment variable

Tools

web_search

Search the web using Perplexity Sonar API for real-time information.

Parameters:

• query (required): The search query or question

• model (optional): Perplexity model - "sonar", "sonar-pro", "sonar-deep-research", "sonar-reasoning", "sonar-reasoning-pro" (default: "sonar")

• recency_filter (optional): Time filter for results (e.g., 'week', 'day')

• city (optional): City name for location-based search

• country (optional): Two letter ISO country code

web_search_academic

Search academic sources using Perplexity Sonar API for scholarly information. Prioritizes peer-reviewed papers and academic journals.

Parameters:

• query (required): The search query or question

• model (optional): Perplexity model - "sonar", "sonar-pro", "sonar-deep-research", "sonar-reasoning", "sonar-reasoning-pro" (default: "sonar")

• recency_filter (optional): Time filter for results (e.g., 'week', 'day')

• city (optional): City name for location-based search

• country (optional): Two letter ISO country code

Configuration

Set your Perplexity API key:

export PERPLEXITY_API_KEY="your-api-key-here"