Research
ChatGPT ads measured: 0.42% of responses — near the ceiling?
ChatGPT ads measured at 0.42% of responses across 8,243 prompts. Google AI Overview ads inside the AI summary at 0.24%. ChatGPT's surface may already be near its ceiling.
Get the real AI Overview UI responses (also called AIO or AIO mode): sources, inline videos, sponsored ads, plus the surrounding organic SERP, People Also Ask, and related searches in one request. Markdown out, any country, any scale.
4.7 on G2
No credit card required.
curl -X POST https://api.cloro.dev/v1/monitor/google \
-H "Authorization: Bearer sk_live_your_api_key_here" \
-H "Content-Type: application/json" \
-d '{
"query": "What do you know about Tesla\'s latest updates?",
"country": "US",
"include": {
"aioverview": {
"markdown": true
}
}
}' {
"success": true,
"result": {
"organicResults": [],
"peopleAlsoAsk": [],
"relatedSearches": [],
"aioverview": {
"text": "...",
"markdown": "...",
"sources": [],
"videos": [],
"ads": []
}
}
} cloro returns the AI Overview answer alongside the surrounding SERP that drives traffic to it. The same key also gets you ChatGPT, Perplexity, Gemini, AI Mode, and Copilot.
AI Overview doesn't render in isolation. It sits at the top of a Google SERP next to People Also Ask, related searches, and the organic results that influence its citations. There is no public AI Overview API, so the only way to see what users see is to capture the whole surface.
Google runs the most aggressive anti-automation stack on the public web. DIY pipelines fail within hours of a SERP UI update, and AI Overview's render path adds another layer of detection on top. cloro absorbs the maintenance so your AI Overview tracking keeps running through every Google revision.
AI Overview only renders for a subset of queries, and for the same prompt, whether it triggers at all shifts by country, session, and Google's current algorithm pass. Cited sources re-rank with every refresh of the underlying SERP. You need repeated sampling per region to see when AI Overview shows up and which sources it picks.
Google ships no AI Overview endpoint. The Search SERP API doesn't include it, and Vertex AI doesn't expose it. To read what real users see, you have to render the SERP and parse it. Surfer's analysis measured ~20% overlap between AI APIs and the rendered UI across LLMs.
AI Overview embeds inline videos and sponsored shopping ads directly inside the AI answer, alongside the cited source cards. None of these special objects lives in any Google API. cloro returns them as structured JSON, plus the surrounding organic SERP, PAA, and related searches.
Parse markdown, sources, inline videos, sponsored ads, plus the surrounding organic SERP, PAA, and related searches from one endpoint.
import requests
response = requests.post(
"https://api.cloro.dev/v1/monitor/google",
headers={
"Authorization": "Bearer sk_live_your_api_key_here",
"Content-Type": "application/json"
},
json={
"query": "What do you know about Tesla's latest updates?",
"country": "US",
"include": {
"aioverview": {
"markdown": True
}
}
}
)
print(response.json()){
"success": true,
"result": {
"organicResults": [
{
"position": 1,
"title": "Tesla Software Update 2024: FSD Beta and Range Improvements",
"link": "https://electrek.co/tesla-software-update-2024",
"displayedLink": "electrek.co",
"snippet": "Tesla's latest software update...",
"page": 1
}
],
"peopleAlsoAsk": [
{
"question": "What are the latest Tesla FSD features?",
"type": "LINK",
"title": "Tesla Full Self-Driving: Latest Features and Improvements",
"link": "https://tesla.com/fsd/features",
"snippet": "Tesla's FSD Beta now ..."
}
],
"relatedSearches": [
{
"query": "Tesla software update 2024.12",
"link": "https://google.com/search?q=tesla+software+update+2024.12"
}
],
"aioverview": {
"text": "Tesla's recent updates...",
"markdown": "### Tesla Recent Updates\n\nTesla's recent updates...",
"sources": [
{
"position": 1,
"url": "https://tesla.com/updates/fsd",
"label": "Tesla FSD Updates",
"description": "Latest Full Self-Driving improvements and capabilities including enhanced visual processing and neural network training"
}
],
"videos": [
{
"url": "https://youtube.com/watch?v=tesla-fsd-v12",
"title": "Tesla FSD v12 Demo: 30 Minutes Hands-Free",
"thumbnail": "https://i.ytimg.com/vi/tesla-fsd-v12/hq.jpg",
"source": "YouTube",
"platform": "youtube",
"date": "3 days ago",
"duration": "32:14"
}
],
"ads": [
{
"position": 1,
"title": "Wall Connector — Tesla",
"url": "https://shop.tesla.com/wall-connector",
"domain": "shop.tesla.com",
"description": "Charge up to 44 mph at home. Compatible with all Tesla vehicles.",
"price": {
"value": "475.00",
"currency": "USD"
},
"store": "Tesla"
}
]
}
}
} Pick a plan that fits your volume. Price per credit drops as you scale.
Credit cost per request varies by provider. The rates below apply to async/batch requests; sync requests add a +2 credit surcharge.
Google News uses the same pricing as Google Search.
The AI Overview block (often abbreviated AIO or AIO mode) renders at position 0 above organic results, marked by a small sparkle/asterisk icon and the words "AI Overview" or "Search Labs · AI Overview" depending on the rollout phase. The logo is Google's standard four-color sparkle (blue, red, yellow, green dots arranged in an asterisk pattern). When scraping programmatically, you don't need to identify it visually — pass `include: {"aioverview": {"markdown": true}}` and the API returns the parsed block in `result.aioverview` if present, or omits the field if AI Overview didn't render for that query.
Google has never shipped one. The Search SERP API returns organic results without the AI Overview block; Vertex AI gives you Gemini without the Google index. AI Overview is a rendered UI feature, not an API surface. To monitor it you have to render the SERP and parse it, which is what cloro does.
Yes, and the share is growing. AI Overview increasingly embeds YouTube videos directly in the answer (with platform, duration, thumbnail) for how-to and product queries, plus sponsored shopping ads (price, store, domain) for commercial intent. cloro returns both as `aioverview.videos` and `aioverview.ads`. If your competitive set is running Performance Max or YouTube ads, this is the surface that places them.
Yes. The same `/v1/monitor/google` request returns organic results, AI Overview (if triggered), People Also Ask, related searches, and sponsored results in one payload. AI Overview citations correlate strongly with organic top-3, so capturing both lets you measure that relationship directly.
Substantially. We see roughly 30–50% trigger-rate divergence between US and EU on the same prompt set. Some EU regions roll out AI Overview slower or scope it tighter under DMA constraints. Pass `country` per request to capture the regional variation.
Each request runs against Google live with typical latency of 30–60 seconds end-to-end and no caching layer. Citations reflect Google's current SERP pass, not yesterday's snapshot.
Use `POST /v1/monitor/google/async` with `aioverview` in the include block. Submit a batch, receive results via webhook. This is the right pattern when you're sampling thousands of prompts per day across multiple countries.
DIY Google SERP scraping is the hardest sustained anti-automation fight on the public web. Realistic in-house cost for AI-Overview-aware monitoring at production volume is $5–10k/month before engineering time, and the success rate is fragile. cloro's Hobby plan ($100/month) covers the same workload with managed reliability.
ChatGPT ads measured at 0.42% of responses across 8,243 prompts. Google AI Overview ads inside the AI summary at 0.24%. ChatGPT's surface may already be near its ceiling.
Scrape the Google AI Overview block in 2026: parse dynamic source citations, handle desktop/mobile layout shifts, and extract embedded sponsored ads.
We tested 12 LLM visibility tracking tools on real brand-monitoring workflows across ChatGPT, Perplexity, Gemini, and Google AI Overview. What works, what doesn't.