Perplexity Advanced MCP

🚀 Perplexity Advanced MCP

The "Perplexity Advanced MCP" toolkit integrates services from OpenRouter and Perplexity to provide enhanced Model Context Protocol (MCP) capabilities. It supports simple queries and complex reasoning, and allows attaching files as context input.

🚀 Quick Start

The "Perplexity Advanced MCP" toolkit offers a convenient way to enhance your model context protocol capabilities. You can quickly install and start it using the following methods.

📦 Installation

Quick Installation via Smithery

One-click installation through the Smithery platform:

npx -y @smithery/cli install @code-yeongyu/perplexity-advanced-mcp --client claude

Quick Startup with uvx

Use the uvx tool to start the MCP server:

uvx perplexity-advanced-mcp -o <openrouter_api_key> # or -p <perplexity_api_key>

You can also configure the API key through environment variables:

export OPENROUTER_API_KEY="your_key_here"
# or
export PERPLEXITY_API_KEY="your_key_here"

uvx perplexity-advanced-mcp

⚠️ Important Note

• Providing both OpenRouter and Perplexity API keys will result in an error.
• When both command-line parameters and environment variables are present, command-line parameters take precedence.

This CLI tool is built on Typer, offering a friendly and intuitive command-line interaction experience.

✨ Features

MCP Search Tool

This package integrates an MCP search tool, implemented through the ask_perplexity function. It supports the following features:

• Simple Queries: Provides quick responses.
• Complex Reasoning: Supports file attachments (XML format required).

📚 Documentation

API Key Configuration

You can choose to configure the API key for OpenRouter or Perplexity in the following ways:

• Command-line parameters: -o <openrouter_api_key> or -p <perplexity_api_key>
• Environment variables: OPENROUTER_API_KEY or PERPLEXITY_API_KEY

Model Mapping

In the src/perplexity_advanced_mcp/config.py file, the mapping between query types and models is as follows:

OpenRouter Models:

• Simple Queries: perplexity/sonar
• Complex Reasoning: perplexity/sonar-reasoning

Perplexity Models:

• Simple Queries: sonar-pro
• Complex Reasoning: sonar-reasoning-pro

🔧 Technical Details

This project was inspired by the exploration of the "vibe coding" trend. With the help of Cline + Cursor IDE, over 95% of the code was generated through voice input (transcribed by Wispr Flow). The project structure and code were all assisted by AI.

From project setup to file organization, all code was automatically generated and reviewed by LLM. Even the GitHub Actions workflow and release process were handled automatically by Cursor. As a human developer, my main tasks were:

1. Start/stop the MCP server to assist AI in thorough testing.
2. Copy and provide error logs for troubleshooting.
3. Provide links to relevant documentation and examples of the Python MCP SDK.
4. Propose modification suggestions for code that did not meet expectations.

📄 License

The development of this project aims to explore the potential of AI tools in the field of software engineering, especially redefining the coding experience through the combination of voice input and LLM. Although it is still in its early stages, we believe this direction will bring a more efficient and natural programming method for developers.

If you are interested in this project or have any questions, please stay tuned for our updates!