# Perplexity

> Search, ask, research, and reason via the official Perplexity MCP.

[Canonical HTML page](https://top-mcps.com/mcp/perplexity) · [server.json](https://top-mcps.com/mcp/perplexity.json) · [methodology](https://top-mcps.com/about/methodology)

## Install

### Claude Desktop — `claude_desktop_config.json`

Paste under mcpServers. Fully quit and reopen Claude after editing.

```json
{
  "mcpServers": {
    "perplexity": {
      "command": "npx",
      "args": [
        "-y",
        "@perplexity-ai/mcp-server"
      ],
      "env": {
        "PERPLEXITY_API_KEY": "${PERPLEXITY_API_KEY}"
      }
    }
  }
}
```

### Claude Code — `CLI or .mcp.json`

Run from your repo. Commit .mcp.json to share with your team.

```shell
# export PERPLEXITY_API_KEY=your_perplexity_key
claude mcp add perplexity -- npx -y @perplexity-ai/mcp-server
```

### Cursor — `.cursor/mcp.json`

Global path: ~/.cursor/mcp.json. Reload window after editing.

```json
{
  "mcpServers": {
    "perplexity": {
      "command": "npx",
      "args": [
        "-y",
        "@perplexity-ai/mcp-server"
      ],
      "env": {
        "PERPLEXITY_API_KEY": "${PERPLEXITY_API_KEY}"
      }
    }
  }
}
```

### VS Code — `.vscode/mcp.json`

VS Code uses the "servers" key (not "mcpServers").

```jsonc
{
  "servers": {
    "perplexity": {
      "command": "npx",
      "args": [
        "-y",
        "@perplexity-ai/mcp-server"
      ],
      "env": {
        "PERPLEXITY_API_KEY": "${PERPLEXITY_API_KEY}"
      }
    }
  }
}
```

### Windsurf — `~/.codeium/windsurf/mcp_config.json`

Open via Cascade → hammer icon → Configure.

```json
{
  "mcpServers": {
    "perplexity": {
      "command": "npx",
      "args": [
        "-y",
        "@perplexity-ai/mcp-server"
      ],
      "env": {
        "PERPLEXITY_API_KEY": "${PERPLEXITY_API_KEY}"
      }
    }
  }
}
```

### Cline — `cline_mcp_settings.json`

Open via the Cline sidebar → MCP Servers → Edit.

```json
{
  "mcpServers": {
    "perplexity": {
      "command": "npx",
      "args": [
        "-y",
        "@perplexity-ai/mcp-server"
      ],
      "env": {
        "PERPLEXITY_API_KEY": "${PERPLEXITY_API_KEY}"
      }
    }
  }
}
```

### Continue — `~/.continue/config.json`

Continue uses modelContextProtocolServers with a transport block.

```json
{
  "experimental": {
    "modelContextProtocolServers": [
      {
        "transport": {
          "type": "stdio",
          "command": "npx",
          "args": [
            "-y",
            "@perplexity-ai/mcp-server"
          ],
          "env": {
            "PERPLEXITY_API_KEY": "${PERPLEXITY_API_KEY}"
          }
        }
      }
    ]
  }
}
```

### Codex CLI — `~/.codex/config.toml`

Codex uses TOML. Each server is a [mcp_servers.<name>] subtable.

```shell
# ~/.codex/config.toml
[mcp_servers.perplexity]
command = "npx"
args = [
  "-y",
  "@perplexity-ai/mcp-server",
]
env = { PERPLEXITY_API_KEY = "${PERPLEXITY_API_KEY}" }
```

### Zed — `~/.config/zed/settings.json`

Zed calls them "context_servers". Settings live-reload on save.

```jsonc
{
  "context_servers": {
    "perplexity": {
      "command": {
        "path": "npx",
        "args": [
          "-y",
          "@perplexity-ai/mcp-server"
        ]
      },
      "env": {
        "PERPLEXITY_API_KEY": "${PERPLEXITY_API_KEY}"
      }
    }
  }
}
```

### ChatGPT — `ChatGPT → Apps directory`

Perplexity doesn't ship a hosted HTTPS endpoint today. ChatGPT supports remote MCP servers only — to use this server in ChatGPT you'll need to deploy it to a public HTTPS URL first (e.g. via Cloudflare Workers or Vercel) or wait for an official remote build.

```none

```

## At a glance

- **Maintainer:** Perplexity
- **Transport:** stdio, Streamable HTTP
- **Auth model:** API key
- **Required secrets:** PERPLEXITY_API_KEY
- **Supported clients:** Claude, Cursor, Any MCP-compatible client, Perplexity API
- **License:** MIT
- **Language:** TypeScript
- **Latest version:** latest
- **Last verified:** 2026-05-27
- **GitHub stars:** 2,233 (fetched 2026-06-02T11:55:52.251Z)
- **Score:** 88/100 (rubric 2026-04 — see https://top-mcps.com/about/methodology)
- **Source:** https://github.com/perplexityai/modelcontextprotocol

## Quick answer

**What it does.** Exposes four Perplexity Sonar capabilities as MCP tools: perplexity_search, perplexity_ask, perplexity_research, perplexity_reason.

**Best for:**
- Cited, source-linked answers
- Deep research (sonar-deep-research)
- Reasoning over retrieved facts
- Conversational follow-up over a topic
- High-quality fact-checking

**Not for:**
- Free-only workflows (paid API credits)
- Use cases that want raw search hits without an answer summary
- Real-time news streaming

## Description

The official Perplexity MCP gives AI agents access to Perplexity's Sonar models: a citation-grounded search, a conversational ask, a deep research mode, and a reasoning mode for complex analysis.

## Why it matters

Perplexity's Sonar models are the most-cited choice when an answer needs source attribution. The official MCP makes that available inside Claude, Cursor, and any MCP host.

## Key features

- Four Sonar modes: search, ask, research, reason
- Cited source URLs in every response
- stdio + HTTP transports
- `PERPLEXITY_API_KEY` auth
- MIT, vendor-maintained

## FAQ

### Which mode should I use?

Use `perplexity_search` for ranked results, `perplexity_ask` for a conversational answer, `perplexity_research` for deep multi-source synthesis, `perplexity_reason` for analytical chains.

### Does it cost?

Yes — Perplexity API credits are consumed per request. Pricing varies by Sonar model tier.

### Is it official?

Yes — maintained by Perplexity at github.com/perplexityai/modelcontextprotocol under MIT.

## Changelog

- **2026-05-27** — Refreshed install snippets and fact sheet; verified for 2026.
- **2025-04-05** — Initial directory listing.