ChatGPT Ads library: how to see competitor ads in 2026

Henry Purchase

Co-Founder

Here is the first thing to know about the ChatGPT Ads Library: it does not exist.

There is no chatgpt.com/ads/library. No transparency center. No public, searchable database where you type a competitor's name and see every ad they are running, the way you can on Meta or Google. OpenAI has not built one, and given how ChatGPT ads work, they may never build one in the form you are expecting.

That is a problem if you are trying to research what your competitors are doing on the fastest-growing ad channel of 2026. It is also an opportunity, because the businesses that figure out how to see inside this channel early get an edge over everyone still waiting for an official tool.

This guide covers why there is no native ChatGPT ad library, what competitor data you can actually capture today, the methods operators are using to build their own libraries, and the tools indexing ChatGPT ads at scale.

Is there a ChatGPT ad library?

No. OpenAI does not publish a public library of ChatGPT ads.

On Meta, every active ad is discoverable in the Meta Ad Library by law. On Google, the Ads Transparency Center shows you any advertiser's running ads. These exist because those ad surfaces are public by default: an ad on Facebook or in Google search results is shown in a shared, visible space.

ChatGPT ads are different at a structural level. Each ad is served privately, inside a one-to-one conversation between a single user and the model. There is no shared feed, no public results page, no archive. OpenAI's own documentation describes the ad as appearing below a relevant response, matched to the context and intent of that specific conversation. The only transparency mechanism is user-facing: you can dismiss an ad, report it, or tap to learn why you were shown it. None of that gives a competitor a way to see what you are running.

So when people search "ChatGPT ad library," they are looking for something that does not exist as an OpenAI product. The rest of this guide is about how to get the intelligence anyway.

Why traditional ad spy tools cannot see ChatGPT ads

The reason SpyFu, SEMrush, and the Meta Ad Library work is that they were built for ad surfaces that are public and predictable. ChatGPT ads break all three assumptions those tools rely on.

The ads are private. Each one is served inside a single conversation. There is no shared surface to scrape, no public archive to index. A tool built to read the Meta Ad Library has nothing to read here.

The ads are contextual, not keyword-triggered. Two people asking the same question can see different ads, because OpenAI matches on the broader topic and intent of the conversation, the user's prior chats, and how they have interacted with ads before. There is no single keyword that reliably surfaces a competitor's ad, so keyword-based spy tools cannot map the space.

The ads are dynamic. Inventory shifts constantly. An ad that ran this morning might be retired by afternoon. A competitor who raised their bid at lunch might dominate every relevant conversation by evening. A static screenshot goes stale within days.

This is why the legacy ad-spy industry has not produced a working ChatGPT solution. The data they are built to consume is not there.

What competitor data you can actually capture

You can still build useful competitive intelligence. The trick is focusing on the signals that are reliably observable. Three are worth capturing systematically.

The sponsored card itself. The headline, body copy, brand name, and click-through URL of any ad you encounter. This is what shows on screen. Capture the full card with a timestamp so you can track how a competitor's creative changes over time.

The triggering intent. Not the user's exact words, but the conversational intent behind the ad. If asking "what is the best CRM for a small agency" surfaces a sponsored card for a particular brand, the intent is "CRM software, SMB, decision stage." Cluster by intent, not exact phrasing, because the platform matches on meaning, not keywords.

The advertiser's frequency and depth. Some brands run a single creative. Some rotate twenty. Some appear on every adjacent intent in a category, others only on bottom-funnel comparison questions. Frequency and depth tell you more about how serious and mature a competitor's account is than any single ad does.

Everything beyond those three is modelling, not measurement. Daily spend estimates, exact CPMs, share-of-voice percentages: none of that is public data on ChatGPT. Any tool claiming to give you a competitor's precise ChatGPT ad spend is estimating, so treat those numbers as directional at best.

Three ways to build a ChatGPT ads library

Method 1: Manual observation

The free method. Build a list of 50 to 100 prompts that map to your category, then run them on a regular schedule and screenshot every sponsored card that appears. Tag each capture with the date, the intent cluster, the advertiser, the creative angle, and the offer.

Two practical notes. Use a fresh ChatGPT session each time so prior ad interactions do not skew what you are served. And you will need a Free or Go tier account, because ads do not appear for Plus or Pro users.

Realistic cost: 4 to 8 hours a week per category. You will capture a small slice of total inventory, but you will get high-quality reads on your direct competitors. This is the right starting point for a small team testing whether the channel matters before paying for tooling.

Method 2: Third-party intelligence tools

A handful of platforms are indexing ChatGPT ads at scale and selling access. The landscape as of mid-2026:

The right pick depends on budget and how many competitors you are tracking. Adthena is built for enterprise procurement and prices that way. Focal includes the Ad Library across every plan because we use it on our own client work daily. Manual observation is the free floor.

Method 3: The hybrid (where most teams land)

Most operators end up combining the two. They use a third-party library for breadth, knowing which 50 brands are active in their category and what new creatives appeared this week, and manual observation for depth, running specific buyer-journey prompts to see what triggers what.

Set up properly, the hybrid takes about 30 minutes a week and produces roughly ten times the coverage of pure manual observation at a fraction of the time cost.

How to actually use a ChatGPT ads library

Capturing ads is the easy part. Turning the data into something your team ships from is where the value is.

Tag every ad by intent and funnel stage. Sort your captures into top of funnel (ads on awareness questions), middle of funnel (ads on comparison questions), and bottom of funnel (ads on decision questions). The winning creative patterns are different at each stage, so this sorting is what makes the data usable.

Look for patterns, not single winners. One competitor running one ad tells you nothing. Four competitors all opening with a price hook tells you the price hook converts. Three competitors all using the same testimonial format tells you that format works. The repetition is the signal. Individual ads are noise.

Brief from the patterns, not the ads. When you build your next ChatGPT ad, reference the pattern, not the specific creative. "We are testing a price-led hook because four competitors are running one" is a brief your team can build from. Copying a competitor's exact ad is theft and converts worse anyway.

Turn patterns into a launch queue. For each pattern worth testing, decide which week it ships, which campaign it slots into, and what success looks like. A library without a launch queue is research, not output. The teams winning ChatGPT ads ship two to four creative variations a week off competitor pattern reads. The teams losing collect screenshots and never launch.

How this works inside Focal

Focal's Ad Library captures live competitor ChatGPT ads, indexes them by industry and creative angle, and pipes the patterns straight into a creative brief, so the gap between "what is my competitor running" and "here is my variation to test" closes inside one tool instead of across a spreadsheet, a screenshot folder, and a separate brief doc.

You filter by industry, brand, or angle. You save the patterns worth testing. You generate variations built for the ChatGPT auction, then push them toward launch. It is the workflow from the section above, without the manual labour.

The Ad Library is one part of the wider platform we are building for ChatGPT advertisers. Join the waitlist here, or if you would rather have it run for you, our ChatGPT Ads agency uses this exact intelligence on client campaigns.

Will OpenAI ever launch an official ChatGPT ad library?

Possibly, but do not wait for it.

Regulatory pressure could force some form of transparency, especially in the EU, where the Digital Services Act already requires large platforms to maintain ad repositories. If ChatGPT ads scale in Europe, an EU-mandated ad archive is plausible. But even if that happens, a compliance-driven repository tends to be minimal: who ran the ad and when, not the rich, intent-tagged, pattern-ready intelligence a performance team actually needs.

The practical takeaway: an official library, if it ever arrives, will likely be a thin transparency tool, not a competitive-intelligence product. The teams building real ChatGPT ad intelligence now will keep their edge regardless.

Frequently asked questions

Is there an official ChatGPT ad library?

No. OpenAI does not publish a public, searchable library of commercial ChatGPT ads. Ads are served privately inside individual conversations, so there is no shared surface to browse the way you can with the Meta Ad Library or Google's Ads Transparency Center. The only native transparency is a user-facing control to dismiss an ad or see why it was shown.

How do I see what ads my competitors are running on ChatGPT?

Three ways: manually, by running category prompts in a Free or Go tier account and screenshotting the sponsored cards; with a third-party tool that indexes ChatGPT ads at scale, like the Focal Ad Library or Adthena; or a hybrid of both. There is no official tool, so all competitor research relies on observation or third-party indexing.

Can I see how much a competitor spends on ChatGPT ads?

No. Spend data is not public on ChatGPT. Any tool quoting a competitor's exact ChatGPT ad spend is estimating from a model, not measuring. You can reliably observe their creative, their offers, and how often and broadly they appear, but not their budget.

Why can't SpyFu or SEMrush track ChatGPT ads?

Because those tools were built for public, keyword-triggered ad surfaces. ChatGPT ads are private, served inside individual conversations, matched on conversational intent rather than keywords, and constantly changing. There is no public data for legacy spy tools to scrape.

Do I need a paid ChatGPT account to see ads?

The opposite. Ads only appear to Free and Go ($8/month) tier users. Plus, Pro, and Business accounts see no ads at all, so you cannot do manual ad research from a Plus or Pro account.

What is the best ChatGPT ad library tool in 2026?

It depends on your needs. For agencies and in-house teams who want competitor ads indexed and pushed into a creative workflow, the Focal Ad Library is built for that. For enterprise brands needing whole-market monitoring, Adthena's ChatGPT Intelligence module is the heavier option. For teams just testing the channel, manual observation in a spreadsheet is a free start.

My question is not listed here.

Ask in the FutureProof community or email support@tryfocal.com.