Perplexity Search
Wire Perplexity models through OpenRouter so your agent can run web-grounded research from a configured .env template.
Overview
perplexity-search is an agent skill most often used in Idea (also Validate competitors, Grow content) that configures OpenRouter access to Perplexity Sonar models for grounded research queries.
Install
npx skills add https://github.com/davila7/claude-code-templates --skill perplexity-searchWhat is this skill?
- OpenRouter-backed Perplexity models: sonar-pro, sonar-pro-search, sonar, sonar-reasoning-pro, sonar-reasoning
- Sonar Pro Search for multi-step agentic search with tool use and intermediate queries
- Configurable DEFAULT_MODEL, DEFAULT_MAX_TOKENS, and DEFAULT_TEMPERATURE in .env
- Model comparison section with best-for traits and example complex research prompts
- Copy-from-template workflow: .env.example to .env with OPENROUTER_API_KEY
- 5+ Perplexity model variants documented via OpenRouter naming
Adoption & trust: 600 installs on skills.sh; 27.8k GitHub stars; 2/3 security scanners passed (skills.sh audits).
What problem does it solve?
You need current, source-backed answers for market or technical questions but only have a generic LLM without search or Perplexity routing set up.
Who is it for?
Builders validating ideas with live web research and choosing between fast Sonar and agentic Sonar Pro Search.
Skip if: Production apps that need a self-hosted search pipeline or compliance-reviewed data sources without third-party API keys.
When should I use this skill?
You need web-grounded research through Perplexity models and must configure OpenRouter credentials and model defaults.
What do I get? / Deliverables
Your agent environment is ready to call the right Perplexity model through OpenRouter with documented defaults for depth, tokens, and temperature.
- Configured .env for Perplexity via OpenRouter
- Model selection guidance for query type
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Research is the primary shelf because the skill documents model selection and API setup for discovery and synthesis, not shipping code. Research subphase fits Sonar Pro Search and comparative deep-dive use cases described in the model comparison guide.
Where it fits
Compare quantum computing vendor approaches before picking a niche product angle.
Draft an article outline with up-to-date sources and model-appropriate depth.
How it compares
API and model routing template, not an MCP server or built-in Claude web search toggle.
Common Questions / FAQ
Who is perplexity-search for?
Solo builders using davila7 Claude Code templates who want Perplexity search models callable via OpenRouter from their agent setup.
When should I use perplexity-search?
Use it in Idea for market and tech research, in Validate for competitor comparisons, or in Grow when drafting content that needs fresh citations.
Is perplexity-search safe to install?
It requires an OpenRouter API key in .env; review the Security Audits panel on this page and never commit secrets to git.
SKILL.md
READMESKILL.md - Perplexity Search
# Perplexity Search Environment Configuration # Copy this file to .env and fill in your actual API key # OpenRouter API Key # Get your key from: https://openrouter.ai/keys OPENROUTER_API_KEY=sk-or-v1-your-api-key-here # Optional: Default model to use # Options: sonar-pro, sonar-pro-search, sonar, sonar-reasoning-pro, sonar-reasoning # DEFAULT_MODEL=sonar-pro # Optional: Default max tokens for responses # DEFAULT_MAX_TOKENS=4000 # Optional: Default temperature (0.0 - 1.0) # DEFAULT_TEMPERATURE=0.2 # Perplexity Model Comparison Guide to different Perplexity models available through OpenRouter and when to use each. ## Available Models All Perplexity models are accessed through OpenRouter using the format: `openrouter/perplexity/[model-name]` ### Sonar Pro Search **Model ID**: `openrouter/perplexity/sonar-pro-search` **Best for:** - Complex multi-step research questions - Queries requiring deep analysis and synthesis - Situations needing comprehensive source exploration - Comparative analyses across multiple domains - Research requiring agentic reasoning workflow **Characteristics:** - Most advanced agentic search system - Executes multi-step reasoning workflows - Uses tools and intermediate queries - Provides most comprehensive answers - Higher cost due to extensive processing **Use cases:** - "Conduct a comprehensive analysis of competing CAR-T cell therapy approaches, including mechanism differences, clinical outcomes, and cost-effectiveness" - "Compare quantum computing approaches for drug discovery with traditional computational methods across multiple metrics" - Research questions requiring synthesis from many sources **Pricing** (approximate): - Input: $3/million tokens - Output: $15/million tokens - Request fee: $18 per 1000 requests **Context window**: 200K tokens ### Sonar Pro **Model ID**: `openrouter/perplexity/sonar-pro` **Best for:** - General-purpose research and search - Balanced performance and cost - Standard scientific queries - Quick information gathering - Most use cases **Characteristics:** - Enhanced capabilities over base Sonar - Good balance of quality and cost - Reliable for most queries - Faster than Pro Search - Recommended default choice **Use cases:** - "What are the latest developments in CRISPR base editing?" - "Summarize recent clinical trials for Alzheimer's treatment" - "Explain how transformer architectures work in modern LLMs" - Standard literature searches - Technical documentation queries **Pricing** (approximate): - Lower cost than Pro Search - Good cost-performance ratio **Context window**: 200K tokens ### Sonar **Model ID**: `openrouter/perplexity/sonar` **Best for:** - Basic searches and queries - Cost-sensitive applications - Simple fact-finding - High-volume queries - Quick lookups **Characteristics:** - Base model with solid performance - Most cost-effective option - Faster response times - Good for straightforward queries - Lower accuracy than Pro variants **Use cases:** - "What is the molecular weight of aspirin?" - "When was CRISPR-Cas9 first used in humans?" - "List the main symptoms of diabetes" - Simple fact verification - Basic information retrieval **Pricing** (approximate): - Lowest cost option - Best for high-volume simple queries **Context window**: 200K tokens ### Sonar Reasoning Pro **Model ID**: `openrouter/perplexity/sonar-reasoning-pro` **Best for:** - Complex logical reasoning tasks - Multi-step problem solving - Technical analysis requiring step-by-step thinking - Mathematical or computational problems - Queries needing explicit reasoning chains **Characteristics:** - Advanced reasoning capabilities - Shows step-by-step thinking - Better for analytical tasks - Excels at technical problem-solving - More structured outputs **Use cases:** - "Walk through the steps to design a clinical trial for testing a novel cancer therapy" - "Analyze the computational complexity of different protein folding algorithms" - "Reason through the molecular