# Perplexity (MCP Server)

> Web search, Q&A, deep research and reasoning powered by Perplexity

- Type: MCP server
- Publisher: Perplexity (official)
- Categories: Research, AI & ML
- Website: https://github.com/perplexityai/modelcontextprotocol
- Source: https://github.com/perplexityai/modelcontextprotocol
- Transport: streamable-http
- Server URL: https://api.perplexity.ai/mcp
- Launch command: npx -y @perplexity-ai/mcp-server
- Sign-in: api-key
- Environment variables: PERPLEXITY_API_KEY
- GitHub stars: 2550
- Listed: 2025-03-10

## About

Perplexity's official MCP server provides direct web search plus conversational answers, deep research and reasoning backed by Perplexity's APIs. It is available as a hosted remote server with an API key bearer token or locally via npx.

## Install

1. **Connect it in ChatGPT** Turn on Developer mode in ChatGPT (Settings → Security and login), open Plugins, select + and paste https://api.perplexity.ai/mcp as the MCP server URL. Add it to a chat from the composer's tools menu.

2. **Or add it to Codex or the API**

   Codex CLI:

   ```bash
   codex mcp add perplexity --url https://api.perplexity.ai/mcp --bearer-token-env-var PERPLEXITY_API_KEY
   ```

   config.toml:

   ```toml
   [mcp_servers.perplexity]
   url = "https://api.perplexity.ai/mcp"
   bearer_token_env_var = "PERPLEXITY_API_KEY"
   ```

   Responses API:

   ```python
   import os
   from openai import OpenAI
   
   client = OpenAI()
   
   response = client.responses.create(
       model="gpt-6-astra",
       tools=[{
           "type": "mcp",
           "server_label": "perplexity",
           "server_url": "https://api.perplexity.ai/mcp",
           "authorization": os.environ["PERPLEXITY_API_KEY"],
           "require_approval": "never",
       }],
       input="Search for the latest benchmarks of open-weight LLMs",
   )
   print(response.output_text)
   ```

## Tools

- `perplexity_search`
- `perplexity_ask`
- `perplexity_research`
- `perplexity_reason`

## Example prompts

- "Search for the latest benchmarks of open-weight LLMs"
- "Do deep research on the US residential solar market"
- "What changed in the latest Python release?"

## Related

- [Browser Use](https://chatgptplugins.directory/mcp/browser-use.md): Run Browser Use cloud browser agents or control a local browser over MCP
- [Memory](https://chatgptplugins.directory/mcp/memory.md): Knowledge graph-based persistent memory across conversations
- [Mem0](https://chatgptplugins.directory/mcp/mem0.md): Hosted long-term memory for AI agents and clients
- [Sequential Thinking](https://chatgptplugins.directory/mcp/sequential-thinking.md): Structured, reflective step-by-step problem solving for LLMs
- [Fetch](https://chatgptplugins.directory/mcp/fetch.md): Fetch web pages and convert them to markdown for LLMs
- [Serena](https://chatgptplugins.directory/mcp/serena.md): Semantic code retrieval and editing toolkit for coding agents

---
Source: https://chatgptplugins.directory/mcp/perplexity · Index: https://chatgptplugins.directory/llms.txt · Independent and community-run. Not affiliated with or endorsed by OpenAI. ChatGPT is a trademark of OpenAI.