Perplexity Search Integration

Programmatic access to Perplexity AI's search API. Execute search queries and get AI-generated answers with citations and related questions.

What This Does

This Actor connects to Perplexity AI's API and processes search queries. For each query, you get back an AI-generated answer, source citations, and optionally related follow-up questions. Think of it as ChatGPT's search feature, but automated and scriptable.

Use Cases

• Research automation: Batch process multiple queries at once
• Content research: Get sourced information for articles or reports
• Competitive analysis: Track how AI surfaces information about topics
• Data pipelines: Integrate AI search into existing workflows
• Question answering systems: Build on top of Perplexity's search capability

Requirements

You need a Perplexity API key. Get one at perplexity.ai by signing up and navigating to API settings.

Input

{
  "apiKey": "pplx-abc123...",
  "queries": [
    "What are the latest developments in quantum computing?",
    "How does Rust's ownership model prevent memory leaks?"
  ],
  "model": "llama-3.1-sonar-small-128k-online",
  "maxTokens": 1024,
  "temperature": 0.2,
  "returnCitations": true,
  "returnRelatedQuestions": true
}

Input Fields

Field	Type	Required	Description
apiKey	string	Yes	Your Perplexity API key
queries	array	Yes	List of search queries to process
model	string	No	Model to use (default: sonar-small)
maxTokens	integer	No	Max response length (default: 1024)
temperature	number	No	Randomness 0.0-2.0 (default: 0.2)
returnCitations	boolean	No	Include sources (default: true)
returnRelatedQuestions	boolean	No	Include follow-ups (default: true)

Available Models

• llama-3.1-sonar-small-128k-online: Fast and cost-effective
• llama-3.1-sonar-large-128k-online: More capable, balanced performance
• llama-3.1-sonar-huge-128k-online: Most accurate, highest quality

Output

{
  "query": "What are the latest developments in quantum computing?",
  "answer": "Recent developments in quantum computing include...",
  "citations": [
    {
      "number": 1,
      "url": "https://example.com/article"
    }
  ],
  "model": "llama-3.1-sonar-small-128k-online",
  "tokensUsed": 847,
  "success": true
}

Output Fields

Field	Type	Description
query	string	Original search query
answer	string	AI-generated response
citations	array	Source URLs referenced
relatedQuestions	array	Suggested follow-up questions
model	string	Model that processed the query
tokensUsed	integer	API tokens consumed
success	boolean	Whether query succeeded
error	string	Error message if failed

Example Queries

Research Multiple Topics

{
  "apiKey": "your-key",
  "queries": [
    "What is the current state of fusion energy research?",
    "How do mRNA vaccines work?",
    "What are the key differences between Redis and Memcached?"
  ]
}

Focused Technical Search

{
  "apiKey": "your-key",
  "queries": ["Explain Rust's borrow checker in simple terms"],
  "temperature": 0.0,
  "maxTokens": 512
}

Creative Exploration

{
  "apiKey": "your-key",
  "queries": ["What are emerging trends in urban architecture?"],
  "temperature": 0.8,
  "returnRelatedQuestions": true
}

Performance

• Query processing: 2-5 seconds per query
• Batch processing: Queries executed sequentially
• Rate limits: Depends on your Perplexity API plan

Error Handling

If a query fails, the Actor continues processing remaining queries. Failed queries return error details in the output:

{
  "query": "failed query",
  "success": false,
  "error": "API Error (429): Rate limit exceeded"
}

Common errors:

• 401: Invalid or missing API key
• 429: Rate limit exceeded
• 500: Perplexity service error

Pricing

This Actor uses the Pay-Per-Event pricing model. You are charged primarily for successful search results.

• Start Fee: $0.05 per 1,000 runs (negligible)
• Result Fee: $1.50 per 1,000 queries ($0.0015 per query)
• Compute: Minimal (covered mostly by the event fees)

Note: You must also pay for your own Perplexity API usage. This Actor charges a small fee for the automation infrastructure and Apify platform features.

Notes

The quality of answers depends on the model selected. The small model is faster and cheaper but less capable than the large or huge models. For most use cases, the small model is sufficient.

Temperature controls randomness: 0.0 is deterministic and focused, higher values produce more creative but potentially less accurate results.

Citations are extracted from Perplexity's response and include source URLs. The actual content snippets from sources are not always available depending on API response format.