perplexity-search

Real-time web searches retrieve current information beyond model knowledge cutoffs by querying live web sources. This Skill performs AI-powered web searches using Perplexity models via OpenRouter, returning grounded answers with source citations suitable for literature discovery, technical research, and fact verification across biomedical and general domains.

Yes. This Skill integrates Perplexity models (sonar, sonar-pro, sonar-reasoning variants) via OpenRouter using LiteLLM as the interface layer. You need an OpenRouter API key, LiteLLM installed, and environment configuration to route requests and access multiple Perplexity models.

Web citations ground search results in source attribution. This Skill returns citations alongside answers from web searches, enabling you to verify information sources and trace reasoning back to specific documents across biomedical, technical, and general domains.

Perplexity offers multiple model variants: sonar for general queries, sonar-pro for balanced performance, and sonar-reasoning-pro for deeper reasoning tasks. This Skill supports switching between these models to optimize for cost, speed, or reasoning depth based on your research needs.

Setup requires setting the OPENROUTER_API_KEY environment variable and installing LiteLLM. This Skill handles credential routing and model selection internally, so once configured, you run queries without managing authentication between requests.

Yes. This Skill performs agentic web searches applicable to real-time information retrieval, literature discovery, and rapid fact verification. It works across biomedical, technical, and general domains, returning usage data alongside citations for tracking and optimization.