One API for AI answers and search results
Each engine returns structured JSON: answer text, citations, brand entities, product cards and ads, with the same request shape and per-call credit pricing across all of them.
Every page below documents one engine end to end: the request options it accepts, the fields it returns, what it costs in credits and a worked example. Start with an assistant, ChatGPT or Gemini, if you are measuring how answers describe, recommend and cite brands, and with Google Search or Google News if you need rankings or headlines in a pipeline. Each engine also has field-level pages (sources, entities, shopping cards, People Also Ask, the local pack) for the parts you will actually store. Terms used throughout are defined in the glossary.
Engines
ChatGPT
The answer ChatGPT shows its users, with every cited source in order, sentence-level citations and its product cards.
PPerplexity
Perplexity's answer with its citations, related queries, the searches it ran, and any videos, images, places, hotels or shopping cards it showed.
GGemini
Gemini's grounded answer with its sources, inline citations and places.
CMicrosoft Copilot
Copilot's Bing-grounded answer with every source it cited.
GGrok
Grok's answer with its cited sources, including X posts, and the searches it ran.
AGoogle AI Mode
Google's conversational AI Mode answer with citation pills, places, shopping cards, inline products, videos and ads.
SGoogle Search
Structured Google results: organic listings, ads, People Also Ask, knowledge graph, local pack and the AI Overview.
NGoogle News
Google News results for any query, country and interface language, up to ten pages deep.