Perplexity Search MCP
Enables AI-powered search capabilities through Perplexity AI's continuously refreshed index. Supports web, academic, and SEC search modes with advanced filtering options including geographic location and customizable result limits.
README
Perplexity Search MCP
This is a tool based on the Model Context Protocol (MCP) that wraps the search functionality of Perplexity AI into a standardized tool that can be called via MCP.
This project is built with TypeScript and runs on the Bun runtime.
✨ Features
- Provides the
perplexity-searchtool via MCP. - Supports various parameters for Perplexity search, such as:
query: A single or multiple search queries.max_results: Specifies the maximum number of results to return.country: Filters search results by geographic location.search_mode: Supportsweb,academic, andsecsearch modes.
- Communicates with the host environment via standard input/output (Stdio).
🚀 Getting Started
1. Prerequisites
- Ensure you have Bun installed.
- You will need an API key from Perplexity AI.
2. Installation
After cloning the project, run the following command in the root directory to install the required dependencies:
bun install
3. Configuration
This server requires the PERPLEXITY_API_KEY environment variable to be set.
4. Running the Server (Development)
To run the server directly from the TypeScript source for development, execute the following command:
bun run index.ts
The server will start and listen for MCP messages on standard I/O.
5. Building the Project
You can also build the project into a single, optimized JavaScript file.
bun run build
This command compiles index.ts and places the output in dist/perplexity-search-mcp.js.
After building, you can run the compiled file directly:
bun dist/perplexity-search-mcp.js
or
node dist/perplexity-search-mcp.js
🛠️ Tool Definition
perplexity-search
Description: Get ranked search results from Perplexity’s continuously refreshed index with advanced filtering and customization options.
Input Parameters:
| Parameter | Type | Description | Default |
|---|---|---|---|
query |
string or string[] |
The search query or queries to execute. | Required |
max_results |
number |
The maximum number of search results to return (max 20). | 10 |
country |
string |
Country code to filter results by geographic location (e.g., 'US', 'GB', 'DE'). | Optional |
max_tokens_per_page |
number |
Controls the maximum number of tokens retrieved from each webpage. Higher values provide more comprehensive content but may increase processing time. | 2048 |
search_mode |
string |
The search mode. Can be web, academic, or sec. |
web |
Output:
The tool returns a JSON string containing an array of search results from the Perplexity API.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。