perplexity-api

Purpose

This skill provides access to Perplexity's Sonar API, allowing real-time web searches, answers with citations, and search-augmented LLM responses to enhance AI-driven applications.

When to Use

Use this skill for tasks requiring current web data, such as answering user queries with verifiable sources, augmenting LLM outputs with live information, or performing research that needs citations. Avoid it for static data or when real-time access isn't needed to prevent unnecessary API calls.

Key Capabilities

• Perform real-time web searches with query parameters for filtering (e.g., by date or domain).
• Retrieve answers as JSON objects including text responses and citation URLs.
• Integrate with LLMs to generate search-augmented responses, combining API results with model inferences.
• Support for multiple query types: standard search, focused queries, and citation extraction.
• Rate limiting awareness, with up to 100 requests per minute per API key.

Usage Patterns

Always set the API key via environment variable before invoking. Use HTTP GET or POST requests for queries, parse JSON responses, and handle asynchronous results if needed. For AI agents, integrate as a subroutine: first check local knowledge, then call this API for web updates. Pattern: Authenticate -> Send query -> Process response -> Log or cache results.