Perplexity Deep Research MCP
# Perplexity Deep Research MCP
[](https://smithery.ai/server/@arjunkmrm/perplexity-deep-research)
A Model Context Protocol (MCP) server for Perplexity AI API with web search capabilities. This server allows AI assistants to perform web searches using Perplexity's chat completions API with the sonar-deep-research model.
## Features
- Provides a `deep_research` tool for AI assistants to perform web searches
- Uses Perplexity's chat completions API with the sonar-deep-research model
## Tool: deep_research
The server provides a `deep_research` tool with the following input parameters:
- `query` (required): The search query to perform
- `search_recency_filter` (optional): Filter search results by recency (options: month, week, day, hour)
## Configuration
### Environment Variables
- `PERPLEXITY_API_KEY`: Your Perplexity API key (required)
## Response Format
The response from the `deep_research` tool includes:
- Content of search results
- Citations for the information
## Installing via Smithery
To install perplexity-deep-research for Claude Desktop automatically via [Smithery](https://smithery.ai/server/@arjunkmrm/perplexity-deep-research):
```bash
npx -y @smithery/cli install @arjunkmrm/perplexity-deep-research --client claude
```
## License
MIT
TDQS
Scored across 1 tool
With only one tool, there is no possibility of ambiguity or overlap between tools. The tool 'deep_research' has a clearly distinct and singular purpose, making it impossible for an agent to misselect among alternatives.
A single tool inherently exhibits perfect naming consistency, as there are no other tools to compare against. The name 'deep_research' follows a clear and descriptive pattern without any deviations or mixed conventions.
A single tool is generally too few for most server purposes, as it limits functionality and flexibility. While the tool is well-described for deep web research, the server's scope feels thin and underdeveloped, which could hinder agent workflows that require more granular operations.
The server is severely incomplete for a research domain, as it only offers a single broad search tool. There are obvious gaps, such as no tools for filtering results, managing queries, accessing specific sources, or performing follow-up actions, which will likely cause agent failures in complex research tasks.