Perplexity AI

AI search engine that provides real-time web-grounded answers with source citations. Unlike traditional search, Perplexity synthesizes information from multiple sources into coherent responses.

Official docs: https://docs.perplexity.ai/

---

When to Use

Use this skill when you need to:

• Get real-time information with source citations
• Research topics that require up-to-date web data
• Answer complex questions that need multi-source synthesis
• Perform academic or SEC filings research
• Get AI-generated summaries grounded in current web content

---

Prerequisites

1. Create an account at https://perplexity.ai/
2. Navigate to API settings: https://perplexity.ai/account/api
3. Generate an API key

Set environment variable:

export PERPLEXITY_API_KEY="pplx-your-api-key"

---

Important: When using $VAR in a command that pipes to another command, wrap the command containing $VAR in bash -c '...'. Due to a Claude Code bug, environment variables are silently cleared when pipes are used directly.

bash -c 'curl -s "https://api.example.com" -H "Authorization: Bearer $API_KEY"' | jq .

How to Use

1. Chat Completions (Recommended)

AI-powered Q&A with web search and citations. OpenAI SDK compatible.

Basic query with sonar model:

Write to /tmp/perplexity_request.json:

{
  "model": "sonar",
  "messages": [
    {"role": "user", "content": "What is the current price of Bitcoin?"}
  ]
}

Then run:

bash -c 'curl -s -X POST "https://api.perplexity.ai/chat/completions" --header "Authorization: Bearer ${PERPLEXITY_API_KEY}" --header "Content-Type: application/json" -d @/tmp/perplexity_request.json' | jq .

With system prompt:

Write to /tmp/perplexity_request.json:

{
  "model": "sonar",
  "messages": [
    {"role": "system", "content": "Be precise and concise. Answer in bullet points."},
    {"role": "user", "content": "What are the latest developments in AI?"}
  ]
}

Then run:

bash -c 'curl -s -X POST "https://api.perplexity.ai/chat/completions" --header "Authorization: Bearer ${PERPLEXITY_API_KEY}" --header "Content-Type: application/json" -d @/tmp/perplexity_request.json' | jq .

Advanced query with sonar-pro:

Write to /tmp/perplexity_request.json:

{
  "model": "sonar-pro",
  "messages": [
    {"role": "user", "content": "Compare the latest iPhone and Samsung Galaxy flagship phones"}
  ],
  "temperature": 0.2,
  "web_search_options": {
    "search_context_size": "high"
  }
}

Then run:

bash -c 'curl -s -X POST "https://api.perplexity.ai/chat/completions" --header "Authorization: Bearer ${PERPLEXITY_API_KEY}" --header "Content-Type: application/json" -d @/tmp/perplexity_request.json' | jq .

2. Search with Domain Filter

Limit search to specific domains.

Write to /tmp/perplexity_request.json:

{
  "model": "sonar",
  "messages": [
    {"role": "user", "content": "Latest research on transformer architectures"}
  ],
  "search_domain_filter": ["arxiv.org", "openai.com", "anthropic.com"]
}

Then run:

bash -c 'curl -s -X POST "https://api.perplexity.ai/chat/completions" --header "Authorization: Bearer ${PERPLEXITY_API_KEY}" --header "Content-Type: application/json" -d @/tmp/perplexity_request.json' | jq .

Exclude domains (add - prefix):

Write to /tmp/perplexity_request.json:

{
  "model": "sonar",
  "messages": [
    {"role": "user", "content": "Best programming practices"}
  ],
  "search_domain_filter": ["-reddit.com", "-quora.com"]
}

Then run:

bash -c 'curl -s -X POST "https://api.perplexity.ai/chat/completions" --header "Authorization: Bearer ${PERPLEXITY_API_KEY}" --header "Content-Type: application/json" -d @/tmp/perplexity_request.json' | jq .

3. Search with Time Filter

Filter by recency.

Write to /tmp/perplexity_request.json:

{
  "model": "sonar",
  "messages": [
    {"role": "user", "content": "Major tech news"}
  ],
  "search_recency_filter": "week"
}

Then run:

bash -c 'curl -s -X POST "https://api.perplexity.ai/chat/completions" --header "Authorization: Bearer ${PERPLEXITY_API_KEY}" --header "Content-Type: application/json" -d @/tmp/perplexity_request.json' | jq .

Filter by date range:

Write to /tmp/perplexity_request.json:

{
  "model": "sonar",
  "messages": [
    {"role": "user", "content": "AI announcements"}
  ],
  "search_after_date_filter": "12/01/2024",
  "search_before_date_filter": "12/31/2024"
}

Then run:

bash -c 'curl -s -X POST "https://api.perplexity.ai/chat/completions" --header "Authorization: Bearer ${PERPLEXITY_API_KEY}" --header "Content-Type: application/json" -d @/tmp/perplexity_request.json' | jq .

4. Academic Search

Search academic sources.

Write to /tmp/perplexity_request.json:

{
  "model": "sonar",
  "messages": [
    {"role": "user", "content": "Recent papers on large language model alignment"}
  ],
  "search_mode": "academic"
}

Then run:

bash -c 'curl -s -X POST "https://api.perplexity.ai/chat/completions" --header "Authorization: Bearer ${PERPLEXITY_API_KEY}" --header "Content-Type: application/json" -d @/tmp/perplexity_request.json' | jq .

5. Raw Search API

Get raw search results without AI synthesis.

Write to /tmp/perplexity_request.json:

{
  "query": "Claude AI Anthropic",
  "max_results": 5
}

Then run:

bash -c 'curl -s -X POST "https://api.perplexity.ai/search" --header "Authorization: Bearer ${PERPLEXITY_API_KEY}" --header "Content-Type: application/json" -d @/tmp/perplexity_request.json' | jq .

With domain and time filters:

Write to /tmp/perplexity_request.json:

{
  "query": "machine learning tutorials",
  "max_results": 10,
  "search_domain_filter": ["github.com", "medium.com"],
  "search_recency_filter": "month",
  "country": "US"
}

Then run:

bash -c 'curl -s -X POST "https://api.perplexity.ai/search" --header "Authorization: Bearer ${PERPLEXITY_API_KEY}" --header "Content-Type: application/json" -d @/tmp/perplexity_request.json' | jq .

6. Deep Research (Long-form)

For comprehensive research reports.

Write to /tmp/perplexity_request.json:

{
  "model": "sonar-deep-research",
  "messages": [
    {"role": "user", "content": "Write a comprehensive analysis of the electric vehicle market in 2024"}
  ],
  "reasoning_effort": "high"
}

Then run:

bash -c 'curl -s -X POST "https://api.perplexity.ai/chat/completions" --header "Authorization: Bearer ${PERPLEXITY_API_KEY}" --header "Content-Type: application/json" -d @/tmp/perplexity_request.json' | jq .

7. Reasoning Model

For complex problem-solving.

Write to /tmp/perplexity_request.json:

{
  "model": "sonar-reasoning-pro",
  "messages": [
    {"role": "user", "content": "Analyze the pros and cons of microservices vs monolithic architecture for a startup"}
  ]
}

Then run:

bash -c 'curl -s -X POST "https://api.perplexity.ai/chat/completions" --header "Authorization: Bearer ${PERPLEXITY_API_KEY}" --header "Content-Type: application/json" -d @/tmp/perplexity_request.json' | jq .

---

Models

Model	Description	Best For
sonar	Lightweight, cost-effective	Quick facts, news, simple Q&A
sonar-pro	Advanced search, deeper understanding	Complex queries, detailed research
sonar-reasoning-pro	Multi-step reasoning with search	Problem-solving, analysis
sonar-deep-research	Exhaustive research, reports	Academic research, market analysis

---

Response Format

Chat completions return:

{
  "id": "uuid",
  "model": "sonar",
  "choices": [
  {
  "message": {
  "role": "assistant",
  "content": "The answer with [1] citations..."
  }
  }
  ],
  "citations": ["https://source1.com", "https://source2.com"],
  "search_results": [
  {
  "title": "Source Title",
  "url": "https://source.com",
  "snippet": "Relevant excerpt..."
  }
  ],
  "usage": {
  "prompt_tokens": 10,
  "completion_tokens": 150,
  "total_tokens": 160
  }
}

---

Guidelines

1. Model Selection: Use sonar for simple queries, sonar-pro for complex ones
2. Cost Optimization: Lower search_context_size reduces costs (low/medium/high)
3. Domain Filters: Limit to 20 domains max; use - prefix to exclude
4. Recency Filters: Options are day, week, month, year
5. Temperature: Lower (0.1-0.3) for factual queries, higher for creative
6. Rate Limits: Vary by tier; check https://docs.perplexity.ai/guides/rate-limits-usage-tiers
7. Citations: Response includes numbered citations [1] that map to citations array