Overview
The Search Perplexity action gives your agent access to real-time web search with AI-powered research capabilities. Perfect for:- Real-time information — access current news, trends, and up-to-date data
- Research tasks — get detailed answers with sources and citations
- Fact checking — verify claims and get authoritative information
Search Perplexity Action
- What it does: Searches the web using Perplexity’s AI models to provide sourced answers to your queries with real-time information.

Inputs
- Query: The search query or question you want answered
- Model: Choose from Perplexity’s model lineup based on your needs
Outputs
- Response: The answer to your query
- Citations: Sources and references used in the response
Query Options
You have two ways to set the query:- Manual Query: Write your search query directly
- AI-Generated Query: Let your agent craft the optimal query based on workflow context
Model Selection
| Model | Description | Strength |
|---|---|---|
| Sonar | Quick research and basic searches | ⭐ |
| Sonar Pro | Better searches with more detailed results | ⭐⭐ |
| Sonar Reasoning | Adds reasoning capabilities to search results | ⭐⭐⭐ |
| Sonar Reasoning Pro | Advanced reasoning with better analysis | ⭐⭐⭐⭐ |
| Sonar Deep Research | Uses Perplexity’s deep research function | ⭐⭐⭐⭐⭐ |
Sonar Reasoning Pro is recommended for most use cases as it provides the best balance of reasoning capability, search quality, and performance.
Best Practices
Query Optimization
Query Optimization
- Be specific about what information you need
- Include timeframes when relevant (“latest”, “2024”, “recent”)
- Use AI query generation when context is important
- Test different phrasings for complex topics
- Include industry or domain context for better results
Model Selection Strategy
Model Selection Strategy
- Start with lower-tier models for simple queries
- Use Deep Research for comprehensive analysis
- Consider cost vs. thoroughness trade-offs
- Use Reasoning models when you need logical connections
- Scale model choice to query complexity