Skip to main content

Overview

Fast Search reduces search latency and cost in agent workflows. Set search_type: "fast" on a Search API request. The Search API charges $1.00 per 1,000 Fast Search requests instead of $5.00 for standard web search.
For Fast Search inside the Agent API, see Web Search: Search Type. Fast Search is a search type, not the Agent API fast preset.
If you omit search_type, the request uses standard web search. Set search_type to "fast" on a POST /search request. Results use the same response format as standard web search. First, install a library and set PERPLEXITY_API_KEY. With Perplexity Python library versions 0.43.4 and 0.43.5, use extra_body as shown below: passing search_type="fast" directly fails the library’s request validation.
Fast Search accepts max_results from 1 to 20. Values above 20 are only supported with search_type: "people". See People Search.

SDK support

The examples accommodate the request definitions in Perplexity Python library 0.43.4 and 0.43.5 and TypeScript SDK 0.38.5.
  • Python: use extra_body={"search_type": "fast"}. The library rejects "fast" when passed directly as search_type.
  • TypeScript: cast the value with as any, as shown, until the SDK updates.

Pricing

Search API billing counts each successful POST /search request, the same as standard web search. In the API console, Fast Search requests appear as Fast Search API Requests Count under Search API usage. There are no additional token-based charges. See Pricing.

Next Steps

Search API Quickstart

Filters, multi-query search, and content controls

API Reference

Full Search API parameters

Pricing

Search and tool pricing