Perplexity Ask MCP Server

Perplexity Ask MCP Server is an MCP server that integrates the Perplexity API platform directly into AI coding assistants and desktop environments. It enables developers, researchers, and knowledge workers to perform live internet searches and structured investigations without leaving their AI chat interfaces. The server connects clients to Perplexity Search and Agent APIs, exposing dedicated tools for real-time web queries, conversational question answering, step-by-step reasoning, and deep multi-step research. By routing requests through Perplexity's curated model presets, it retrieves ranked web search results complete with metadata, filters by domain or recency, and streams research progress for lengthy analytical tasks. Developers working in coding environments like Cursor, VS Code, Windsurf, and Claude Desktop use this server to retrieve current documentation, verify library updates, debug emerging framework errors, and investigate technical topics using up-to-date online information. The server can be accessed remotely through Perplexity's hosted streamable HTTP endpoint or run locally via Node.js using the official npm package. It also accommodates corporate enterprise proxies and custom network configurations, providing flexible access to web-grounded LLM intelligence across diverse developer workflows.

Category: AI & LLM Tooling

Tags: information-retrieval, perplexity, research, web search

Visit Perplexity Ask MCP Server

How to install and configure Perplexity Ask MCP Server

  1. Obtain an API key from the Perplexity API Portal. 2. To connect via the hosted remote server in Claude Code, run: claude mcp add --transport http perplexity https://api.perplexity.ai/mcp --header "Authorization: Bearer YOUR_API_KEY" 3. To run locally in Claude Code, execute: claude mcp add perplexity --env PERPLEXITY_API_KEY="your_key_here" -- npx -y @perplexity-ai/mcp-server 4. For manual setup in clients like Cursor (~/.cursor/mcp.json) or Claude Desktop (claude_desktop_config.json), configure npx with args ["-y", "@perplexity-ai/mcp-server"] and pass PERPLEXITY_API_KEY under env. 5. Alternatively, launch a self-hosted HTTP server using Docker: docker run -p 8080:8080 -e PERPLEXITY_API_KEY=your_key_here perplexity-mcp-server

What you can do with Perplexity Ask MCP Server

  • Querying current technical documentation and recent software releases through live Perplexity web search with specific domain and recency filters. * Conducting deep multi-step research on complex industry topics using streamed Agent API runs that return comprehensive analytical reports. * Solving complex programming and logical reasoning problems step-by-step directly inside code editors using the specialized reasoning preset. * Asking conversational coding questions grounded in live web information to quickly troubleshoot unfamiliar error codes and deprecated packages. * Executing corporate-compliant research behind organizational proxies by routing outbound requests through dedicated enterprise proxy environment variables.

Key facts

  • https://github.com/perplexityai/modelcontextprotocol
  • AI & LLM Tooling, Web Search & Research
  • information-retrieval, perplexity, research, web search

Part of MCP Servers

Related MCP servers

  • MCP Memory Dashboard — MCP Memory Dashboard is an MCP server desktop interface that connects to the MCP Memory Service to provide visual semantic…
  • MCP Manager — MCP Manager is an MCP server management tool that connects directly to your Claude Desktop environment, enabling users to discover,…
  • MCP Lab — MCP Lab is an MCP server development environment designed for building, testing, and debugging custom Model Context Protocol servers integrated…
  • MCP LLM Integration Server — MCP LLM Integration Server is an MCP server that connects local Large Language Model runtimes with Model Context Protocol clients…
  • MCP NPX Fetch — MCP NPX Fetch is an MCP server that retrieves online resources and transforms web content into structured formats including HTML,…
  • MCP Naver News — MCP Naver News is an MCP server that connects AI assistants to the Naver News API, enabling automated search and…

How do I install Perplexity Ask MCP Server?

You can connect directly to Perplexity's hosted remote endpoint using HTTP transport and your API key. Alternatively, run the local package via npx by specifying @perplexity-ai/mcp-server in your client configuration file, or run the claude mcp add command in Claude Code with your Perplexity API key.

What can Perplexity Ask MCP Server do?

The server provides four core tools: perplexity_search for ranked search results with recency and domain filters, perplexity_ask for rapid conversational answers, perplexity_research for multi-minute deep analytical reports with progress streaming, and perplexity_reason for advanced step-by-step problem-solving using Perplexity Agent API presets.

Which MCP clients work with Perplexity Ask MCP Server?

The server works with MCP-compliant desktop and command-line clients including Cursor, VS Code, Claude Code, Claude Desktop, Kiro, Codex, and Windsurf. Clients connect either through local stdio configurations using Node.js and npx, or over streamable HTTP directly to the remote hosted endpoint or Docker container.

Can I run Perplexity Ask MCP Server behind a corporate proxy?

Yes, the server supports corporate networks and firewalls. You can configure network routing by setting the PERPLEXITY_PROXY environment variable with your proxy host, port, and credentials. The server also checks standard HTTPS_PROXY and HTTP_PROXY variables before attempting direct internet connections.

  • AI Tools
  • Categories
  • Industries
  • CLI Coding Agents
  • MCP Servers
  • MCP Categories