Install guide and config at curatedmcp.com

Perplexity MCP: Ground Your AI Agent in Real-Time Web Research with Citations

Building AI agents that hallucinate facts is the problem Perplexity MCP solves. Instead of letting Claude or Cursor make up answers, this server gives your AI agent direct access to Perplexity's search engine—returning synthesized, cited answers backed by real sources.

What It Does

Perplexity MCP bridges the gap between raw web search and intelligent synthesis. When you connect it to Claude or Cursor, your agent gains the ability to answer questions by actually researching the web, not guessing.

Key capabilities:

• Real-time web grounding — answers pull from current information, not training data
• Source citations — every answer includes direct URLs, so you know where claims come from
• Academic mode — tap scholarly papers and research journals for rigorous answers
• Pro search — multi-step reasoning for complex questions requiring synthesis across multiple sources
• Image and news search — specialized modes for visual research or breaking news
• Follow-up question support — agents can drill deeper without losing context

This is fundamentally different from passing raw search results to your agent. Perplexity doesn't return a list of ten links—it returns a coherent answer built from those sources. For research, fact-checking, market analysis, or any task where you need current, verifiable information, this matters.

How to Install

npx -y @perplexity/mcp

Add to your Claude Desktop configuration:

{
  "mcpServers": {
    "perplexity-mcp": {
      "command": "npx -y @perplexity/mcp",
      "env": {
        "PERPLEXITY_API_KEY": "your-api-key-here"
      }
    }
  }
}

Get your free API key at Perplexity's developer portal. Free tier includes 5 requests daily; Pro plans start at $20/month for unlimited searches.

Real-World Use Cases

• Bug research while debugging — Agent searches GitHub issues, Stack Overflow, and recent discussions to find solutions without you context-switching out of your IDE
• Market research reports — Build competitive analysis by having Claude research current pricing, feature sets, and recent announcements across multiple competitors, with citations for your findings
• Fact-checking claims in code reviews — When reviewing PRs that reference performance benchmarks or library capabilities, your agent can verify claims against current documentation and recent blog posts

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.