Search

Ask any question across ChatGPT, Claude, Perplexity, and Gemini and get each model’s answer plus the sources it cited, straight from your agent.

Overview

The workhorse tool. Your agent passes a question and gets back every model’s prose answer with the sources each cited, in one structured result, so it can reason over what AI search actually says, not just one model’s take.

Example

Ask your agent in plain English and it calls the tool for you:

# you say
What are people saying are the best AI blog tools in 2026?

The tool returns structured JSON to your agent (you never see this directly):

{
  "query": "best AI blog tool 2026",
  "results": [
    { "model": "chatgpt", "answer": "Several tools stand out…", "citations": ["…"] }
    // …one entry per model queried
  ],
  "billing": { "cost": { "usd": "0.32" } }
}

Your agent reads that JSON and answers you in plain English. That reply is written by your own AI; the JSON is just what it reasons over:

# your AI replies
Across the four models, the names that keep coming up are Jasper, Copy.ai, and Writesonic.
ChatGPT and Claude both lead with Jasper; Perplexity leans on independent reviews and cites 3 sources the others don’t.

Full input and output schema: POST /v1/search. Pass an optional models list to narrow which models run (see per-model pricing).