Ask ChatGPT, Claude, Gemini and Perplexity through DataForSEO

Embed

<a href="https://seolyticize.com/docs/ref-get-llm-responses"><img src="https://seolyticize.com/docs/_api/public/orgs/71cb2b7a-a719-4003-b3ea-6321dc1db8ab/spaces/3bf2dbf7-3faa-473c-ba3b-30588f5f8de4/pages/221aea2e-e222-459a-a468-43017509150a/social-card.svg?v=3e7ae33eec8b" alt="Ask ChatGPT, Claude, Gemini and Perplexity through DataForSEO" width="600" height="315"></a>

POSThttps://seolyticize.com/api/v1/tools/get_llm_responses

Asks one question (up to 500 characters) live on the chosen models through DataForSEO — chat_gpt, claude, gemini, perplexity, and chat_gpt_app: the ChatGPT app itself as a person sees it, with the sources it cited and the searches it ran (about $0.004) — with web search on by default, and returns each answer's text, the pages it cited, the follow-up searches it ran (fanOutQueries, where the platform reports them), and which of the client's tracked brands it named and whether the client was named and cited. Needs no assistant keys of the agency's own; nothing is stored. Use it to explore a question before tracking it. Answers 'what do the big models say when asked this' and 'what does ChatGPT search for to answer it'. A paid lookup on the client's account unless anyone asked the same thing recently (served says which); quote it with estimate_cost. (one purchase per model; about $0.01–0.10 each with web search)

Parameters

Idempotency-Keystringheader

Any unique string, up to 200 characters. A retry with the same key and body is answered from the first answer (Idempotent-Replayed: true) for a day, so it never buys or writes twice; the same key with another body is refused.

Request body

application/json

clientIdstringrequired

promptstringrequired

platformsstring[]

= ["chat_gpt","perplexity"]

webSearchboolean

= true

countrystring

Where the web search runs from (ChatGPT, Claude and Perplexity's Sonar models).

= null

modelsobject

A model per platform; defaults gpt-5-mini, claude-sonnet-4-6, gemini-2.5-flash, sonar.

= {}

Show child parameters

chat_gptstring

claudestring

geministring

perplexitystring

maxCostMicrosinteger

Optional spending limit for this call, in micro-USD (1,000,000 = $1). A purchase that would take the call past it is refused before anything is bought. Pass the estimate from estimate_cost, or what the user agreed to spend.

dryRunboolean

True to buy nothing and get the quote instead: what the call would cost at most and what it would buy.

{
  "clientId": "string",
  "prompt": "string",
  "platforms": [
    "chat_gpt",
    "perplexity"
  ],
  "webSearch": true,
  "country": null,
  "models": {},
  "maxCostMicros": 0,
  "dryRun": true
}

Code samples

cURL

curl -X POST "https://seolyticize.com/api/v1/tools/get_llm_responses" \
  -H "Authorization: Bearer $SEOLYTICIZE_API_KEY" \
  -H "Idempotency-Key: <value>" \
  -H "Content-Type: application/json" \
  -d '{"clientId":"string","prompt":"string","platforms":["chat_gpt","perplexity"],"webSearch":true,"country":null,"models":{},"maxCostMicros":0,"dryRun":true}'

TypeScript

Python

PHP

Ruby

Java

Responses

200 The tool's result under data, and its meta.

Example

{
  "data": {},
  "meta": {
    "requestId": "string",
    "costMicros": 0,
    "cached": true,
    "cachedFrom": "string",
    "dataAsOf": "string",
    "sources": [
      "string"
    ],
    "partial": true,
    "next": [
      {
        "tool": "string",
        "args": {},
        "why": "string"
      }
    ],
    "nextCursor": "string",
    "dryRun": true
  }
}

Schema

datarequired

metaobjectrequired

Beside every answer: what it cost, whether it came from cache, how fresh it is, whether it is complete, and what to call next.

Show child parameters

requestIdstringrequired

costMicrosintegerrequired

Charged by this call, micro-USD (1,000,000 = $1).

cachedbooleanrequired

True when nothing was bought.

cachedFromstring

When the oldest cached answer used was bought.

dataAsOfstring

sourcesstring[]

partialbooleanrequired

True when the data is incomplete: a running crawl, unfinished days, unread pages.

nextobject[]required

Show child parameters

toolstring

argsobject

whystring

nextCursorstring

Pass as cursor for the next page.

dryRunboolean

Headers

X-RateLimit-Limitinteger

Requests this key may make per window.

X-RateLimit-Remaininginteger

Requests left in the current window.

X-RateLimit-Resetinteger

Seconds until the current window resets.

X-Request-Idstring

The id to quote about this request.

400 The body was not JSON or did not match the tool's input schema.

Example

{
  "error": "string",
  "code": "string",
  "next": "string",
  "requestId": "string"
}

Schema

errorstringrequired

A sentence written for the caller. Internal failures are reduced to 'The request could not be completed.' and never carry detail.

codestring

A stable snake_case reason to branch on; GET /api/v1/errors lists them all.

nextstring

The call or step that fixes it, when there is one.

requestIdstring

Quote it when asking about a failure.

Headers

X-RateLimit-Limitinteger

Requests this key may make per window.

X-RateLimit-Remaininginteger

Requests left in the current window.

X-RateLimit-Resetinteger

Seconds until the current window resets.

401 No key, or a key that is unknown or revoked.

Same as 400 · POST /api/v1/tools/get_llm_responses

Headers

WWW-Authenticatestring

403 The key's role does not permit the tool, or the key may not reach the named client.

Same as 400 · POST /api/v1/tools/get_llm_responses

Headers

X-RateLimit-Limitinteger

Requests this key may make per window.

X-RateLimit-Remaininginteger

Requests left in the current window.

X-RateLimit-Resetinteger

Seconds until the current window resets.

404 No tool by that name.

Same as 400 · POST /api/v1/tools/get_llm_responses

Headers

X-RateLimit-Limitinteger

Requests this key may make per window.

X-RateLimit-Remaininginteger

Requests left in the current window.

X-RateLimit-Resetinteger

Seconds until the current window resets.

409 A request with this Idempotency-Key is still running.

Same as 400 · POST /api/v1/tools/get_llm_responses

Headers

X-RateLimit-Limitinteger

Requests this key may make per window.

X-RateLimit-Remaininginteger

Requests left in the current window.

X-RateLimit-Resetinteger

Seconds until the current window resets.

422 A well-formed request the product declined, with the reason — a duplicate name, a plan limit, nothing configured to run.

Same as 400 · POST /api/v1/tools/get_llm_responses

Headers

X-RateLimit-Limitinteger

Requests this key may make per window.

X-RateLimit-Remaininginteger

Requests left in the current window.

X-RateLimit-Resetinteger

Seconds until the current window resets.

429 This key has used its window.

Same as 400 · POST /api/v1/tools/get_llm_responses

Headers

X-RateLimit-Limitinteger

Requests this key may make per window.

X-RateLimit-Remaininginteger

Requests left in the current window.

X-RateLimit-Resetinteger

Seconds until the current window resets.

Retry-Afterinteger

Seconds to wait before retrying.

500 The request could not be completed.

Same as 400 · POST /api/v1/tools/get_llm_responses

Headers

X-RateLimit-Limitinteger

Requests this key may make per window.

X-RateLimit-Remaininginteger

Requests left in the current window.

X-RateLimit-Resetinteger

Seconds until the current window resets.

Updated Nov 4, 2026

Was this page helpful?