Perplexity Search

Perplexity Search is a dedicated web search API for retrieving information from the internet. Unlike Sonar Search, which gives Perplexity's Sonar models built-in search capabilities, the dedicated Search API can be used as an independent workflow tool and its results can be passed to any model, agent, or later workflow step.

In Fetch Hive, Perplexity Search gives workflows a standalone search step for finding current information from the web. This makes it useful when you want to separate information retrieval from AI generation and control which model processes the search results afterwards.

Use Cases

Model-independent web research

Search the web using Perplexity and then pass the results to whichever AI model you want. This allows you to use Perplexity for retrieval without requiring a Perplexity model to generate the final response.

Research pipelines

Use Perplexity Search near the beginning of a workflow to collect information before passing the results through summarisation, analysis, classification, extraction, or verification steps.

Multi-model workflows

Retrieve information once and use the same search results across different models or agents. This can be useful when comparing model outputs or assigning different stages of a workflow to different providers.

Why use Perplexity Search with Fetch Hive?

Perplexity Search separates web retrieval from model generation. Instead of tying search to a particular LLM, Fetch Hive can retrieve search results and make that information available to any compatible model, agent, or workflow step.

This gives you more control over how research workflows are structured. Fetch Hive also records the trace behind each run, allowing you to see when search happened, how the results were processed, which models were used afterwards, and how much the overall workflow cost.

FAQ

What can Perplexity Search do in Fetch Hive?

Perplexity Search can retrieve current web information as a dedicated workflow step, with the results available to later workflow steps.

What is the difference between Perplexity Search and Sonar Search?

Sonar Search enables web search directly inside Perplexity Sonar model calls. Perplexity Search is a dedicated search endpoint that can retrieve information independently and pass the results to any model or workflow step.

Can I use Perplexity Search with non-Perplexity models?

Yes. Because search is performed as a separate workflow step, the resulting information can be passed to other supported models and providers.

Can Perplexity Search be used with Fetch Hive agents?

Yes. Search results can become inputs to agents or other workflow steps for further reasoning, analysis, or processing.

Can Fetch Hive track Perplexity Search costs?

Yes. Fetch Hive tracks workflow costs and traces the steps involved in retrieving and processing search results.