Skip to main content
Glama
fastmcp-me

perplexity-mcp

by fastmcp-me
README.md
[![Add to Cursor](https://fastmcp.me/badges/cursor_dark.svg)](https://fastmcp.me/MCP/Details/560/perplexity)
[![Add to VS Code](https://fastmcp.me/badges/vscode_dark.svg)](https://fastmcp.me/MCP/Details/560/perplexity)
[![Add to Claude](https://fastmcp.me/badges/claude_dark.svg)](https://fastmcp.me/MCP/Details/560/perplexity)
[![Add to ChatGPT](https://fastmcp.me/badges/chatgpt_dark.svg)](https://fastmcp.me/MCP/Details/560/perplexity)
[![Add to Codex](https://fastmcp.me/badges/codex_dark.svg)](https://fastmcp.me/MCP/Details/560/perplexity)
[![Add to Gemini](https://fastmcp.me/badges/gemini_dark.svg)](https://fastmcp.me/MCP/Details/560/perplexity)

# perplexity-mcp MCP server

[![smithery badge](https://smithery.ai/badge/perplexity-mcp)](https://smithery.ai/server/perplexity-mcp)

A Model Context Protocol (MCP) server that provides web search functionality using [Perplexity AI's](https://www.perplexity.ai/) API. Works with the [Anthropic](https://www.anthropic.com/news/model-context-protocol) Claude desktop client.

## Example

Let's you use prompts like, "Search the web to find out what's new at Anthropic in the past week."

## Glama Scores

<a href="https://glama.ai/mcp/servers/ebg0za4hn9"><img width="380" height="200" src="https://glama.ai/mcp/servers/ebg0za4hn9/badge" alt="Perplexity Server MCP server" /></a>

## Components

### Prompts

The server provides a single prompt:

- perplexity_search_web: Search the web using Perplexity AI
  - Required "query" argument for the search query
  - Optional "recency" argument to filter results by time period:
    - 'day': last 24 hours
    - 'week': last 7 days
    - 'month': last 30 days (default)
    - 'year': last 365 days
  - Uses Perplexity's API to perform web searches

### Tools

The server implements one tool:

- perplexity_search_web: Search the web using Perplexity AI
  - Takes "query" as a required string argument
  - Optional "recency" parameter to filter results (day/week/month/year)
  - Returns search results from Perplexity's API

## Installation

### Installing via Smithery

To install Perplexity MCP for Claude Desktop automatically via [Smithery](https://smithery.ai/server/perplexity-mcp):

```bash
npx -y @smithery/cli install perplexity-mcp --client claude
```

### Requires [UV](https://github.com/astral-sh/uv) (Fast Python package and project manager)

If uv isn't installed.

```bash
# Using Homebrew on macOS
brew install uv
```

or

```bash
# On macOS and Linux.
curl -LsSf https://astral.sh/uv/install.sh | sh

# On Windows.
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
```

### Environment Variables

The following environment variable is required in your claude_desktop_config.json. You can obtain an API key from [Perplexity](https://perplexity.ai)

- `PERPLEXITY_API_KEY`: Your Perplexity AI API key

Optional environment variables:

- `PERPLEXITY_MODEL`: The Perplexity model to use (defaults to "sonar" if not specified)

  Available models:

  - `sonar-deep-research`: 128k context - Enhanced research capabilities
  - `sonar-reasoning-pro`: 128k context - Advanced reasoning with professional focus
  - `sonar-reasoning`: 128k context - Enhanced reasoning capabilities
  - `sonar-pro`: 200k context - Professional grade model
  - `sonar`: 128k context - Default model
  - `r1-1776`: 128k context - Alternative architecture

And updated list of models is avaiable (here)[https://docs.perplexity.ai/guides/model-cards]

### Cursor & Claude Desktop Installation

Add this tool as a mcp server by editing the Cursor/Claude config file.

```json
  "perplexity-mcp": {
    "env": {
      "PERPLEXITY_API_KEY": "XXXXXXXXXXXXXXXXXXXX",
      "PERPLEXITY_MODEL": "sonar"
    },
    "command": "uvx",
    "args": [
      "perplexity-mcp"
    ]
  }
```

#### Cursor
- On MacOS: `/Users/your-username/.cursor/mcp.json`
- On Windows: `C:\Users\your-username\.cursor\mcp.json`

If everything is working correctly, you should now be able to call the tool from Cursor.
<img width="800" alt="mcp_screenshot" src="https://github.com/user-attachments/assets/4b59774f-646c-41b3-9886-5cfd4c4ca051" width=600>

#### Claude Desktop
- On MacOS: `~/Library/Application\ Support/Claude/claude_desktop_config.json`
- On Windows: `%APPDATA%/Claude/claude_desktop_config.json`

To verify the server is working. Open the Claude client and use a prompt like "search the web for news about openai in the past week". You should see an alert box open to confirm tool usage. Click "Allow for this chat".

  <img width="600" alt="mcp_screenshot" src="https://github.com/user-attachments/assets/922d8f6a-8c9a-4978-8be6-788e70b4d049" />

TDQS

B3.2/5.0

Scored across 1 tool

Disambiguation5/5

There is only one tool, so no ambiguity exists. The agent will always select this tool for web search tasks.

Naming Consistency5/5

The single tool name follows a clear verb_noun pattern (perplexity_search_web), which is consistent and intuitive.

Tool Count3/5

Having only one tool is minimal but acceptable for a narrowly focused 'search' server; however, it feels thin and could benefit from additional related tools.

Completeness2/5

The server only provides web search with recency filtering, lacking features like search result extraction, multiple search modes, or context handling, which limits its usefulness for comprehensive tasks.

Maintenance

ActivityInactive
ResponsivenessNo issues