perplexity-search

Search the web for current scientific answers with source citations.

21|2|Updated Dec 8, 2025
One-click install
npx skills add https://github.com/silverstein/claude-scientific-skills-desktop --skill perplexity-search-silverstein
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: perplexity-search
Source: https://github.com/silverstein/claude-scientific-skills-desktop/tree/main/corpus/perplexity-search
Command: npx skills add https://github.com/silverstein/claude-scientific-skills-desktop --skill perplexity-search-silverstein

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires litellm, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

It solves the problem of answering time-sensitive scientific and technical questions with up-to-date, web-grounded information and source citations beyond an AI model’s knowledge cutoff.

Core Features & Use Cases

  • Real-time web search with citations: Produces grounded answers with references to support verification.
  • Multi-model access via OpenRouter: Supports several Perplexity model variants (e.g., Sonar Pro, Sonar Pro Search, and reasoning-focused options) through a single workflow.
  • Scenarios: Use it to find recent scientific literature, track developments since 2024, validate claims, and summarize emerging trends across domains like biomedical research, ML, and clinical science.

Quick Start

Use the perplexity-search skill to answer a current question with citations by searching for: "What are the latest developments in CRISPR base editing in 2024, and what key studies support those claims?"

Frequently Asked Questions about perplexity-search

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get real-time web search answers with citations for recent scientific literature?

Real-time web search answers with citations are generated by routing queries through Perplexity models via OpenRouter and LiteLLM, grounding responses in current web sources to support literature discovery and fact verification.

Can I use LiteLLM to access Perplexity models like Sonar Pro for current research?

Yes, LiteLLM integrates with OpenRouter to access Perplexity model variants including Sonar Pro and reasoning-focused options, requiring the OPENROUTER_API_KEY configuration to route scientific and technical queries successfully.

What is the best way to track scientific developments published after a model's training cutoff?

Tracking post-cutoff scientific developments is best achieved by querying web-grounded search models that retrieve real-time information, ensuring answers reflect the latest publications and emerging trends across domains.

Does this web search approach support adjusting max_tokens and temperature for research queries?

Yes, the web search workflow supports model selection and allows configuration of max_tokens and temperature parameters, enabling control over response length and creativity for comparative research and trend tracking.

Why do I need an OpenRouter API key to run web-grounded searches?

An OpenRouter API key is required because LiteLLM uses it to authenticate and route requests to Perplexity models, which perform the real-time web searches and return cited answers for current scientific questions.

When should I use real-time web search instead of a standard language model for technical questions?

Real-time web search should be used for time-sensitive technical questions, claim validation, and literature discovery when answers must reflect information published after a standard model's knowledge cutoff, ensuring factual accuracy.