MCP 服务器

Perplexity MCP Server

Connects AI assistants to Perplexity AI's search API, enabling natural language web search with fine-grained control over recency, citations, images, and model parameters directly from MCP clients like Claude.

README

<h1 align="center">Perplexity MCP Server</h1>

<p align="center"> Search the web from Claude Desktop, Claude Code, or any MCP client using Perplexity AI —<br> with fine-grained control over recency, citations, images, and model parameters. </p>

What This Does

This MCP server connects AI assistants to Perplexity AI's search API. Ask questions in natural language and get grounded, cited answers from the live web — directly inside Claude or any MCP-compatible client.

One tool, full control: perplexity_search_web exposes the complete Perplexity API — recency filtering, model selection, temperature, top_k/top_p, citation/image toggles, and streaming.

Getting Started

1. Get a Perplexity API Key

2. Install & Configure

Claude Code (Recommended)

claude mcp add perplexity -- npx -y @jschuller/perplexity-mcp

Then set your API key:

export PERPLEXITY_API_KEY=pplx-your-key-here

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "perplexity": {
      "command": "npx",
      "args": ["-y", "@jschuller/perplexity-mcp"],
      "env": {
        "PERPLEXITY_API_KEY": "pplx-your-key-here"
      }
    }
  }
}

Config location:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

From Source

git clone https://github.com/jschuller/perplexity-mcp.git
cd perplexity-mcp
npm install && npm run build

3. Verify

Ask Claude: "Search the web for the latest developments in quantum computing from the last week"

Tool Reference

`perplexity_search_web`

Parameter	Type	Default	Description
`query`	string	(required)	Search query
`recency`	`day` \| `week` \| `month` \| `year`	`month`	Filter results by time period
`model`	string	`sonar`	Perplexity model (model cards)
`temperature`	number	—	Randomness (0 = deterministic, 2 = creative)
`max_tokens`	integer	—	Maximum tokens to generate
`top_k`	integer	—	Limit high-probability token pool (0 = disable)
`top_p`	number	—	Nucleus sampling threshold
`frequency_penalty`	number	—	Penalize repeated tokens
`presence_penalty`	number	—	Encourage topic variety
`return_citations`	boolean	`true`	Include source citations
`return_images`	boolean	`false`	Include relevant images
`stream`	boolean	`false`	Stream response incrementally

Environment Variables

Variable	Required	Default	Description
`PERPLEXITY_API_KEY`	Yes	—	Your Perplexity API key
`PERPLEXITY_MODEL`	No	`sonar`	Default model for all queries

Available Models

sonar — Standard model (default)
sonar-pro — Enhanced capabilities
See full list: Perplexity Model Cards

Security

API key stays in your local environment — never sent anywhere except the Perplexity API
The server communicates only with api.perplexity.ai over HTTPS
No data is stored or logged beyond the API request lifecycle
See SECURITY.md for vulnerability reporting

Troubleshooting

Issue	Fix
`PERPLEXITY_API_KEY is required`	Set the env var in your MCP client config or shell
`400 invalid_request_error`	Update to v2.1.0+ (fixes JSON Schema validation with Claude Code)
Server not found	Verify `npx @jschuller/perplexity-mcp` runs without error
Connection timeout	Check internet connectivity and Perplexity API status