perplexity

# Perplexity AI Search ## Overview This skill provides access to the Perplexity API for web-grounded AI search and research. It combines the power of large language models with real-time web search, providing accurate, up-to-date answers with source citations. ## When to Use Perplexity vs. Built-in Search **Use Perplexity when:** - You need **current information** (news, prices, events, recent developments) - The user asks for **source citations** or references - Complex **multi-step reasoning** is required - The user specifically mentions Perplexity or wants research-style answers - You need **exhaustive analysis** across multiple sources **Use built-in web search when:** - Simple factual queries - Quick information lookup - The user doesn't need AI-generated synthesis - Basic URL or content retrieval is sufficient ## Model Selection Guide Choose the right model based on task complexity: ### 🔍 Search Models (Quick Facts) Use for simple, factual queries where speed matters. - `sonar` - Default search model with web access. Best for most queries. - `sonar-pro` - Advanced search with deeper understanding. ### 🧠 Reasoning Models (Complex Analysis) Use for complex, multi-step tasks requiring logical thinking. - `sonar-reasoning` - Complex reasoning with web search. - `sonar-reasoning-pro` - Advanced reasoning with deeper content understanding. ### 📚 Research Models (Exhaustive Analysis) Use for comprehensive, in-depth research across multiple sources. - `sonar-research` - Comprehensive research with in-depth analysis. - `sonar-research-pro` - Advanced research with exhaustive analysis and detailed reports. ## Quick Start ### Basic Search ```bash # Simple query (uses sonar by default) scripts/perplexity_search.sh "What is the capital of Germany?" # With custom model scripts/perplexity_search.sh "Latest AI developments" -m sonar-pro # Markdown format with citations scripts/perplexity_search.sh "Tesla stock analysis" -f markdown ``` ### Advanced Research ```bash # Deep research with comprehensive analysis scripts/perplexity_search.sh "Market analysis for electric vehicles in 2025" \ -m sonar-research-pro -c high -f markdown # Pro search mode (multi-step reasoning) scripts/perplexity_search.sh "Compare AI models performance benchmarks" \ -m sonar-reasoning-pro -p pro -f markdown # With custom system prompt scripts/perplexity_search.sh "Analyze tech trends" \ -s "You are a technology analyst. Focus on business implications and market trends." ``` ## Search Context Size Control how much web information is retrieved: - **low** - Faster, fewer sources. Good for simple queries. - **medium** (default) - Balanced. Good for most use cases. - **high** - Most comprehensive. Best for research and detailed analysis. ## Pro Search Mode Available for `sonar-pro` and reasoning models. Controls multi-step tool usage: - **fast** (default) - Standard single-step search. - **pro** - Automated multi-step reasoning with multiple web searches. - **auto** - Automatic classification based on query complexity. ## Setup Requirements ### API Key Configuration The skill requires Perplexity API key. There are two ways to configure it: **Option 1: Skill-specific config file (recommended)** Create `config.json` in the skill directory: ```json { "apiKey": "pplx-your-key-here" } ``` **Option 2: Environment variable** ```bash export PERPLEXITY_API_KEY="your-key-here" ``` To set it permanently (add to ~/.bashrc or ~/.zshrc): ```bash echo 'export PERPLEXITY_API_KEY="your-key-here"' >> ~/.bashrc source ~/.bashrc ``` **Priority:** Config file takes precedence over environment variable. ### Dependencies The script uses bash and curl. Both are typically pre-installed on Linux systems. ## Usage Patterns ### News and Current Events ```bash scripts/perplexity_search.sh "Latest news about AI regulation in Europe" -m sonar ``` ### Competitive Analysis ```bash scripts/perplexity_search.sh "Compare iPhone 15 vs Samsung Galaxy S24 features" \ -m sonar-reasoning-pro -c high -f markdown ``` ### Market Research ```bash scripts/perplexity_search.sh "Electric vehicle market forecast 2025-2030" \ -m sonar-research-pro -c high -p pro -f markdown ``` ### Technical Questions with Current Data ```bash scripts/perplexity_search.sh "Latest Python frameworks for web development 2025" \ -m sonar-reasoning -c medium ``` ## Output Formats - **text** (default) - Plain text with citation references [1], [2], etc. - **markdown** - Markdown formatted response with source links - **json** - Raw API response in JSON format ## Cost Awareness Perplexity API is not free. Be mindful of usage costs: - **Simple queries**: ~$0.005–$0.015 per query - **Deep research**: ~$0.015–$0.03+ per query - **Pro users get $5/month credits** in Perplexity Pro subscription Use reasoning/research models judiciously. Default to `sonar` for most queries. ## List Available Models ```bash scripts/perplexity_search.sh --list-models ``` ## Troubleshooting **Error: PERPLEXITY_API_KEY environment variable not set** - Set up API key as described in "Setup Requirements" above **Error: curl command not found** - Install curl: `apt install curl` or equivalent for your system **Error: Unexpected API response** - Check your API key is valid and has not been revoked - Verify your Perplexity account has API access ## Resources ### scripts/ - **perplexity_search.sh** - Main script for Perplexity API interactions - Supports all Perplexity models - Handles API key discovery from environment or config - Provides multiple output formats - Uses curl for API calls (no Python dependencies) --- **Note:** This skill uses external API calls. Be mindful of rate limits and costs. The API key should never be committed to version control or shared publicly.