HTML pages — the engine's own screen for every answer

Set include.html to true and the task returns a link to the screen the engine showed for the answer, saved as a self-contained HTML page you can put in a report.

Request

curl -X POST https://api.querying.ai/v1/async/task \
  -H "Authorization: Bearer $QUERYING_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "taskType": "GOOGLE", "payload": { "query": "best CRM for a small team", "country": "US", "include": { "html": true } } }'

Supported engines

taskType	The page
`CHATGPT`	The whole ChatGPT screen (sidebar, question, answer and message box) as ChatGPT showed it when the answer finished. Scripts are removed and images are included, so it opens without loading anything.
`GEMINI`	The Gemini screen as Gemini showed it when the answer finished. Scripts are removed and styles, fonts and images are included, so it opens without loading anything.
`PERPLEXITY`	The Perplexity screen as Perplexity showed it when the answer finished, prepared the same way as `GEMINI`.
`META_AI`	The Meta AI screen as Meta AI showed it when the answer finished, prepared the same way as `PERPLEXITY`.
`GOOGLE`	The results page Google returned, with its scripts removed so it opens safely. Google’s styles and embedded thumbnails are kept, and its links open on Google.
`AIMODE`	The AI Mode page Google returned, prepared the same way as `GOOGLE`.

The other engines reject the flag with 422 EVIDENCE_UNSUPPORTED. GET /v1/capabilities marks the supported engines with evidence: true. CHATGPT rejects it with the same error when the task has a location, asks for CN or HK, or continues a conversation with continueFrom: ChatGPT draws its page only for a new conversation answered without a device location.

Response

The task’s response gains an html object. For example:

{
  "html": {
    "status": "ok",
    "url": "https://api.querying.ai/v1/evidence/1791551389-5cf17d26396946be8d68bf514e1dca18.html",
    "expiresAt": "2026-11-08T13:09:49.000Z"
  }
}

Field	Description
`status`	`ok`, or `failed` when no page could be saved.
`url`	The page. Needs no API key.
`expiresAt`	When the URL stops working, 30 days after the task finished.
`error`	`EVIDENCE_FAILED` when `status` is `failed`.

Access and retention

The URL needs no API key, so it works in a link or a report. Anyone with the URL can open it until expiresAt.
After expiresAt the URL answers 404 and the page is deleted. Download what you need to keep.
One IP address can download up to 600 pages a minute; above that it gets 429.

Credits

On CHATGPT, PERPLEXITY and META_AI the page costs 0.25 credits on top of the engine’s price. They are held at submit and charged only when html.status is ok. On GEMINI, GOOGLE and AIMODE the page costs nothing extra.

When the page fails

The task still completes with its answer, and html comes back as { "status": "failed", "error": "EVIDENCE_FAILED" }, when:

the engine’s screen could not be collected, or did not match the answer;
the page could not be stored;
on GEMINI, the task has a location or asks for CN, DE, GB or IR. These always fail.

Resubmit the task to try again.

Errors

Status	Code	When
`422`	`EVIDENCE_UNSUPPORTED`	The engine cannot return an HTML page, or a `CHATGPT` task has a `location`, `CN`, `HK` or `continueFrom`.
`400`	`VALIDATION_ERROR`	`include.html` is not a boolean, or the request sets `include.screenshot: true`. Screenshots are no longer available; use `include.html`.

Markdown and raw responses
ChatGPT · Gemini · Perplexity · Meta AI · Google AI Overviews · Google AI Mode

Get started

Core concepts

Answer engines

Search and web data

Request options

Research

GEO monitoring

Account and data

HTML pages — the engine's own screen for every answer

Request

Supported engines

Response

Access and retention

Credits

When the page fails

Errors

​Request

​Supported engines

​Response

​Access and retention

​Credits

​When the page fails

​Errors

​Related

Request

Supported engines

Response

Access and retention

Credits

When the page fails

Errors

Related