Modelcontextprotocol MCP Server
Perplexity API access to real-time web search and research tools for AI assistants.
Data last scanned last week · Reviewed 2 months ago
Overview
This implementation provides AI agents and assistants with structured access to the Perplexity API platform. It enables the use of the perplexity_search tool, which performs live web searches to retrieve ranked results and associated metadata for up-to-date information gathering.
Our verdict
With a security score of 58/100, this server is classified as high-risk. This is primarily due to the fact that it is a community-maintained project rather than an official vendor release, alongside the lack of a verified audit for its authentication and read-only handling. Developers should exercise caution when using this in environments requiring strict security guarantees.
- Supports active web searching via Perplexity Search API
- Contains only five direct dependencies
- Maintains an active development cycle with recent commits
- Large community footprint with over 2,400 GitHub stars
- Authentication methods have not yet been reviewed
- Read-only capabilities remain unverified
- Maintained by the community instead of the official vendor
Setup
{
"mcpServers": {
"modelcontextprotocol": {
"args": [
"-y",
"@perplexity-ai/mcp-server"
],
"command": "npx"
}
}
}Paste into ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows), then restart Claude Desktop. Some servers require API keys or tokens as environment variables — check the project's README and add an "env" object if needed.
- No read-only mode — running this grants full read/write access.
- No built-in authentication — anyone with access to this config can invoke every tool it exposes.
Tools
| Tool | Description | Risk |
|---|---|---|
| perplexity_search | Direct web search using the Perplexity Search API. Returns ranked search results with metadata, perfect for finding current information. | low |
| perplexity_ask | General-purpose conversational AI with real-time web search using the sonar-pro model. Great for quick questions and everyday searches. | low |
| perplexity_research | Deep, comprehensive research using the sonar-deep-research model. Ideal for thorough analysis and detailed reports. | low |
| perplexity_reason | Advanced reasoning and problem-solving using the sonar-reasoning-pro model. Perfect for complex analytical tasks. | low |
Compatibility
| Client | Local | Docker | Remote | Read-only |
|---|---|---|---|---|
| Claude Desktop | ||||
| Cursor | ||||
| VS Code | ||||
| Windsurf | ||||
| ChatGPT |
Frequently asked questions
›What prerequisites are needed to run this server?
You must obtain a Perplexity API key from the official Perplexity API portal. Additionally, the environment requires Node.js installed to execute the server via npx.
›Can I adjust the output format for the reasoning and research tools?
Yes, you can set the optional parameter 'strip_thinking' to 'true' in the perplexity_reason and perplexity_research tools to remove <think> tags from responses.
›How can I increase the default response timeout?
You can modify the server timeout by setting the 'PERPLEXITY_TIMEOUT_MS' environment variable, which defaults to 300,000 milliseconds.
›How do I configure the server for a corporate proxy environment?
You can route traffic through a company proxy by setting the 'PERPLEXITY_PROXY' environment variable to your proxy host and port.
›Does this server support custom API endpoints?
Yes, you can specify a custom base URL by setting the 'PERPLEXITY_BASE_URL' environment variable if you are not using the default API address.
›Is it possible to adjust the verbosity of server logs?
Yes, you can change the log detail level by setting 'PERPLEXITY_LOG_LEVEL' to DEBUG, INFO, WARN, or ERROR.
Badge
Maintain this server? Add the live badge to your README.