OpenAI started showing ads in ChatGPT in Europe in August 2026 and, a week later, the European Commission designated ChatGPT a very large online search engine (VLOSE). That requires it to publish an ad repository similar to Google's and Meta's. As of 12 November 2026 we have not found an official public repository: here is what is known, what the law requires and how to prepare to access the data once it is published.
Dates from OpenAI and European Commission announcements.
| Date | Milestone |
|---|---|
| Feb 2026 | First ads in ChatGPT, in the US and on the Free and Go plans. |
| 24 Aug 2026 | Ads in 31 European markets: the 27 EU countries plus Iceland, Liechtenstein, Norway and Switzerland. |
| 31 Aug 2026 | The Commission designates ChatGPT a VLOSE (and Reddit and Roblox VLOPs). OpenAI reported 159.1 million monthly EU users of search. |
| ≈ Jan 2027 | End of the four-month window to comply with the stricter obligations, including the Article 39 repository. |
| 12 Nov 2026 | We have not found an official public repository from OpenAI. |
Official source: Commission press release of 31 August 2026. The Commission supervises ChatGPT together with the Irish coordinator (Coimisiún na Meán).
The law sets it for every VLOSE that shows ads. User conversations cannot appear: the repository must not contain personal data of the people who saw the ad.
| Field | What it will have to show | What is not known yet |
|---|---|---|
| Content | The ad and the advertised product, service or brand. | How its creatives will be published in the repository. |
| Advertiser and payer | Identity of the advertiser and, if different, of whoever paid for the ad. | — |
| Period | First and last date shown. | — |
| Targeting | Whether it targeted groups and the main parameters, including exclusion criteria. | How selection by conversation context will be described. |
| Reach | Total recipients reached and, where applicable, a breakdown by Member State of the targeted groups. | The granularity (ranges or figures). |
Public, searchable and with an API. The DSA requires the repository to allow multi-criteria searches, offer API access and keep each ad until one year after it was last shown.
Watch your competitors where data already exists. Brands buying ChatGPT ads are usually on Google, Bing and Meta too. Their libraries can already be extracted.
Track when your brand appears in AI answers. A ChatGPT mention of a brand is not an ad: ads are labelled as such. Mentions in ChatGPT, Gemini or Perplexity, and the sources they cite, can be measured too.
We have not found an official repository yet (12 November 2026). It must meet this obligation within four months of being notified of its VLOSE designation (31 August 2026); the Commission places the deadline in January 2027.
The Commission classified it as a VLOSE because it can search the web to answer user prompts.
In Europe, since 24 August 2026, in 31 markets: the 27 EU countries, Iceland, Liechtenstein, Norway and Switzerland.
No. Article 39 forbids including personal data of the people who saw the ad. The ad, advertiser, payer, dates, targeting and reach will be published.
Article 39 requires API access, but OpenAI has not announced its technical details yet.
The DSA requires keeping each ad until one year after it was last shown. It is not confirmed how ads from before the obligation applies will be handled.
We are tracking it. Leave your email below and we will let you know when there is data to extract.
We will write when OpenAI publishes its repository and we can extract it.