One API returns the parsed ads array on every ChatGPT response — brand, creative, deep-link, and country. Built for ad-ops, brand teams, and ad-tech vendors who need the raw signal, not a packaged dashboard.
4.8 · 42 reviews
cloro returns the parsed ads array on every ChatGPT response. The same API key also gets you Google AI Overview ads, plus organic citations across every major AI engine.
Our 2026-05-26 measurement put ChatGPT ad penetration at 26.5% of responses overall and 49.1% within the US, up from 0.42% three weeks earlier. The competitive-intelligence problem is no longer theoretical.
From 0.42% in April-May to 26.5% overall and 49.1% in the US on 2026-05-26. A competitor buying placements on your priority queries today shows up on ~1 in 2 US responses. Full breakdown.
OpenAI's API returns model output, not the rendered UI. The `ads[]` array, the `bzrcdn.openai.com` creative CDN, the `utm_source=chatgpt.com` tagging — none of it surfaces through any official endpoint. cloro renders and parses the UI for you.
ChatGPT's anti-automation hardens faster than most teams can keep up with. DIY proxy fleets and fingerprint rotation spend more on infra than the data is worth. cloro absorbs that maintenance so your monitoring keeps running.
ChatGPT ads are an English-speaking-market product, and commercial-intent queries surface them at far higher rates. Single-prompt monitoring tells you nothing — you need a query-set × country grid.
The ads[] array is a non-empty array when ChatGPT served a paid placement. Each entry has the brand, creative, and deep-link already parsed.
from cloro import Cloro
client = Cloro(api_key="sk_live_your_api_key_here")
response = client.monitor.chatgpt(
prompt="best HR software for a 500-person company",
country="US",
include={"markdown": True},
)
ads = response.get("result", {}).get("ads") or []
for ad in ads:
print(f"{ad['brand']['name']}: {ad['cards'][0]['title']}"){
"success": true,
"result": {
"text": "For a 500-person company, the leading options include...",
"model": "gpt-5-3",
"sources": [
{
"position": 1,
"url": "https://rippling.com/blog/hr-software-500-employees",
"label": "HR Software for Mid-Market Companies",
"description": "How to choose an HR platform..."
}
],
"ads": [
{
"brand": {
"name": "Rippling",
"url": "https://rippling.com",
"favicon": "https://bzrcdn.openai.com/rippling/favicon.png"
},
"cards": [
{
"title": "Run HR, IT, and Finance in One Platform",
"body": "All your employee data in one system. Free up your team to focus on what matters.",
"url": "https://rippling.com/?utm_source=chatgpt.com&utm_medium=src",
"image": "https://bzrcdn.openai.com/rippling/card.png"
}
]
}
],
"searchQueries": [
"best HR software mid-market",
"HR platform 500 employees"
]
}
}Start free. Price per credit drops as your volume grows — see every tier below.
| Plan | Price / mo | Credits | Per 1k | Concurrency | Seats | Support |
|---|---|---|---|---|---|---|
| Free | $0 | 500 | — | 1 | 1 | Community |
| Lite | $30 | 37,500 | $0.80 | 10 | Unlimited | |
| Hobby | $100 | 250,000 | $0.40 | 20 | Unlimited | |
| StarterPopular | $250 | 650,000 | $0.39 | 50 | Unlimited | |
| Growth | $500 | 1,350,000 | $0.37 | 75 | Unlimited | Priority email |
| Business | $1,000 | 2,800,000 | $0.36 | 100 | Unlimited | Priority email |
| Enterprise 2K | $2,000 | 5,871,025 | $0.34 | 135 | Unlimited | Slack |
| Enterprise 3K | $3,000 | 9,306,606 | $0.32 | 175 | Unlimited | Slack |
| Enterprise 4K | $4,000 | 12,756,261 | $0.31 | 215 | Unlimited | Slack |
| Enterprise 5K | $5,000 | 16,391,783 | $0.31 | 255 | Unlimited | Slack |
| Enterprise | $5,000+ | Increased concurrency, overages on credits and credit discounts for annual contracts.Know more | ||||
Run your category-defining queries through `POST /v1/monitor/chatgpt` and inspect the `ads[]` array on each response. Brand name, ad creative, deep-link, and country are all returned as structured fields. Aggregate across your priority query set on a fixed cadence (weekly is typical) and you have your category's share-of-voice signal. The technical monitoring guide walks through a working Python implementation.
Same endpoint, different lens. `/chatgpt/` is positioned around the full UI surface — citations, sources, shopping cards, entities, brand mentions. `/chatgpt-ads/` is positioned around the `ads[]` array specifically: who's buying placements, what they're saying, and how often. Same `POST /v1/monitor/chatgpt` request; the response payload carries both surfaces, and you parse the one you care about.
Three independent signals, any one of which is sufficient. First, the placement appears in `response.result.ads[]` rather than `sources[]` (OpenAI maintains a strict separation). Second, the creative assets are served from `bzrcdn.openai.com`, OpenAI's brand/creative CDN used exclusively for paid placements. Third, the destination URL is tagged `utm_source=chatgpt.com&utm_medium=src` for advertiser attribution. cloro checks all three for redundancy.
Yes. Define a fixed query set (typically 50-200 priority prompts representing your category), run it on cadence, and aggregate the `ads[]` arrays into a per-brand placement count. Share of voice is straightforward arithmetic from there. The new-advertiser signal — brands that appear for the first time vs your trailing-30-day baseline — is often the highest-signal event in the dataset.
As of our 2026-05-26 measurement: US (49.1%), Canada (33.6%), Australia (19.8%), UK (1.5%), and roughly zero everywhere else. ChatGPT advertising in May 2026 is an English-speaking-market product. If you're tracking non-US/CA/AU/UK markets, expect sparse ad data and budget your monitoring volume accordingly.
On the post-spike rate (~25% overall, ~49% US), daily cadence on a US-focused query set surfaces enough signal to track per-advertiser trends. Weekly is fine for total-volume monitoring and non-US coverage. Hourly polling burns API calls without surfacing proportionally more data — there's no benefit until you're tracking a specific competitive event in real time.
Use the async task API (`POST /v1/async/task` with `taskType: "CHATGPT"`). Submit your priority query set as a batch, receive results via webhook when each completes. Same response shape, same `ads[]` detection logic. Sync concurrency limits depend on your plan; async has none.
DIY ChatGPT ad monitoring requires rotating residential proxies, headless browser fleets, fingerprint rotation, and engineering time to chase ChatGPT's anti-automation cadence. Realistic in-house cost lands at $3-8k/month on infra alone before headcount. cloro's Hobby plan ($100/month) covers 250k requests with managed reliability and zero infra to run.
ChatGPT ads over the last 7 days measure 51.0% of US responses, 53.6% CA, 49.8% AU, and — new since May — 18.8% JP. Everywhere else remains near zero. The surface reverted then re-spiked in June and now looks structural, not transient.
ChatGPT ads render as a `response.result.ads[]` array on the response, served from `bzrcdn.openai.com` and tagged `utm_source=chatgpt.com&utm_medium=src`. Here's how to detect them with cloro's API and build a real monitoring pipeline.
ChatGPT ads are sold direct through OpenAI's sales team at a $200K minimum commitment, with CPC bids of $3-5 general and $8-18 for software and finance. No self-serve manager exists yet. Here's how to actually get placements and what to do if you can't.