AI Assistants

Google AI Overview API

Read the AI Overview Google shows for a query — answer text and cited sources — from a dedicated endpoint. 30 credits flat.

Start freeGET https://app.irmu.com/api/crawl30 credits

Why

Why teams scrape Google AI Overviews

Google AI Overviews holds some of the most commercially useful public data on the web, and almost none of it is available through a stable official feed at the volume real products need. Rate limits, partner-only access and shifting page structures mean teams either give up on the source or maintain a fragile scraper that breaks every few weeks.

AI Overviews sit above the organic results, so for many queries they are the first answer a searcher reads. Irmu's GET /google-ai-overview searches the Google you choose, presses past the consent dialog, expands the overview and returns its text and the pages it cites — no selectors to maintain.

Irmu removes the infrastructure question entirely. One authenticated request handles proxy selection, browser rendering, anti-bot challenges and retries, and AI Extract turns the resulting page into validated JSON against a schema you define. You get query, domain, url, ai_overview, sources, status, credits_charged without writing or maintaining a single selector, and the same endpoint, quota and observability model you already use for every other source.

Use cases

What it powers

AI Overview tracking
See which of your keywords trigger an overview and whether you are cited.
Citation share
Measure how often your pages appear in sources against competitors.
Market comparison
Ask google.de, google.es and google.co.uk the same question.
Content gaps
Find the sources Google trusts for your topics and what they say.

Data

Fields available

Everything below comes back typed and validated against your schema.

query

The search, as it was asked

domain

Which Google was asked (com, de, co.uk…)

url

The results page that was read

ai_overview

The overview's answer as text, or null when Google showed none

sources

Pages the overview cites, with title and url

status

done, queued, running or failed

credits_charged

30, or 0 when the crawl failed

Request

One call, no selectors

request.sh

curl -G https://app.irmu.com/api/crawl \
  -H "Authorization: Bearer $IRMU_API_KEY" \
  --data-urlencode "url=https://www.google.co.uk/search?q=best+web+scraping+api" \
  --data-urlencode "premium=true" \
  --data-urlencode "ai_query=Extract query, domain, url, ai_overview, sources, status as JSON"

Response

Validated JSON

response.json

{
  "url": "https://www.google.co.uk/search?q=best+web+scraping+api",
  "domain": "www.google.co.uk",
  "premium": true,
  "js": true,
  "ai_query": "Extract query, domain, url, ai_overview, sources, status as JSON",
  "credits_charged": 30,
  "status": "completed",
  "data": {
    "query": "best web scraping api",
    "domain": "co.uk",
    "ai_overview": "Popular web scraping APIs include…",
    "sources": [
      {
        "title": "Irmu",
        "url": "https://irmu.com/products/scraping-api"
      }
    ],
    "status": "done"
  },
  "error": null
}

SDKs

Copy, paste, ship

import os, requests

res = requests.get(
    "https://app.irmu.com/api/crawl",
    params={
        "url": "https://www.google.co.uk/search?q=best+web+scraping+api",
        "premium": "true",
        "ai_query": "Extract query, domain, url, ai_overview as JSON",
    },
    headers={"Authorization": f"Bearer {os.environ['IRMU_API_KEY']}"},
)

print(res.json()["data"])

Pricing

Simple credits

Per AI Overview page

30 credits

Requests we fail to serve for internal reasons, or to domains we do not cover, are not billed.

See full pricing →

FAQ