I pointed ChatGPT Work at the Meta ads library, left my desk for half an hour, and came back to a research dashboard with almost 70 notes on how my competitors actually sell. The Meta Ad Library is the free, public record of every ad running on Facebook and Instagram, so it’s the first place to look before you spend anything on ads of your own. Below I’ll show you how to search it yourself, and then how to hand the whole job to ChatGPT Work and let it do the reading for you.

What the Meta Ad Library is

The Meta Ad Library is a searchable archive of the ads currently running across Facebook, Instagram, Messenger and the Audience Network. Meta launched it in 2019 as a transparency tool, and anyone can open it without logging in or paying. If a company is advertising on Meta, you can see its ads.

For any advertiser, you can pull up every active ad, the creative it uses, the date each one started running and the platforms it’s showing on. You can search by keyword to see everyone in a category at once, or go straight to a competitor by name. Meta’s own transparency centre lists it as a research tool, and for marketers that’s a fair description. It shows you what your market is testing right now, paid for by the people doing the testing.

Search it by hand in two minutes

Open the library, set your country, and type in a keyword or a competitor’s name. Then a few filters do most of the work:

• Ad category. Leave it on “All ads” for commercial research. The other categories cover politics, housing, employment and credit ads, which carry extra disclosures.

• Platform. Narrow it to Facebook or Instagram if you only care about one.

• Media type. Filter to images or video, depending on what you’re studying.

• Active status and dates. Filter by when the ads were running.

There’s one trick worth knowing, which Jon Loomer’s guide to the library covers well: Filter for ads that have been running for a long time. Advertisers stop paying for ads that don’t work, so the ones still running after months are your best guess at what’s converting.

What to read in a competitor’s ad, and what you can’t see

To understand a competitor’s ad properly, you need to read three things together: The creative itself, the copy above and below it, and the landing page it clicks through to. Most of the useful detail sits on that landing page, because that’s where the offer, the price and the promise are spelled out. The ad’s job is to get the click, so it’s the page that tells you how they’re actually selling.

It’s also worth knowing what the library leaves out, because that changes how you use it. For ordinary commercial ads you can see the creative and the dates, but the spend, the targeting and the performance stay hidden. Those numbers only appear for political and social-issue ads, and in more detail for ads shown in the EU under the Digital Services Act. So what you’re really studying is your competitors’ messaging, and you infer how well it works from which ads they keep paying to run.

Hand the whole job to ChatGPT Work

Checking one competitor by hand only takes a few minutes. Doing it properly for ten, opening every ad, following each one to its landing page and writing it all up, takes an afternoon, and that’s the work worth delegating.

I started our market research skill inside ChatGPT Work and gave it the campaign context: We were preparing a Black Friday push for our AI Operator course, aimed at professionals and founders in the US, UK, Ireland, Australia and New Zealand. It asked me a few questions about targeting, then opened its own browser and went to the Meta Ad Library by itself.

From there it worked the way a human researcher would. It found the pages it judged to be our competitors, clicked through the ads they were running and followed each ad to its landing page, noting the messaging on both. The whole run took over half an hour. That might sound slow, but I spent that half hour on something else entirely, which is the point of handing it over.

When I came back, there was a full research dashboard waiting. It had an overview with almost 70 substantive buyer observations and 95 broader ones, plus a separate in-depth analysis for each competitor that I could click into to see their whole approach. The library itself has been open to everyone since 2019. What’s changed is that an agent can now read all of it for you in one go.

How it reads video ads

Static ads are easy for an agent to understand, because it can see the image, read the copy around it and open the landing page. Video takes more work, so the skill does two things to turn each video ad into something it can read.

First, it exports one frame for every second of the video. Laid out in order, those frames show the visual story from the opening shot to the end card. Then it transcribes the audio, so it has the exact words being spoken. Put the frames and the transcript together and it knows what the video shows and says, which means a competitor’s video ads go into the same written analysis as their static ones.

Turn competitor findings into your own ads

Once you have the dashboard, the useful work is reading across competitors for patterns: The offers that keep coming up, the formats nobody’s using, the promises everyone leans on. Buffer’s rundown of ways to use the library makes a good checklist for this, from spotting gaps in the market to studying the long-running ads.

For us, those findings become the brief for what we make next. The angles that keep appearing tell us which messages to test first, and the gaps show us where we can say something our competitors don’t. From there it’s a short step to writing the ad copy and building the campaign, which is what the wider AI ad creative system is for. The library is the input that makes the rest of the work specific.

Get the research skill

The market research skill I used here, the one that drives the browser and builds the dashboard, is free, along with the rest of the system. All the prompts and files are in the free 9x community, where our operators share the skills and workflows they actually run. Join, grab the skill and point it at your own market.

Common questions

Is the Meta Ad Library free? Yes. The Meta Ad Library is free and public, with no login required. Anyone can search every active ad running on Facebook, Instagram, Messenger and the Audience Network.

Can I see how much a competitor is spending on Facebook ads? Not for ordinary commercial ads. The library shows the creative, the platforms and the start date, but spend, targeting and performance only appear for political and social-issue ads, and in more detail for ads shown in the EU under the Digital Services Act.

What’s the best way to find a competitor’s best-performing ads? Filter for the ads that have been running longest. Advertisers stop paying for ads that don’t work, so the ones still live after months are your best signal of what’s converting.

Can AI do competitor research in the Meta Ad Library for me? Yes. ChatGPT Work can open its own browser, visit the Meta Ad Library, click through competitors’ ads, follow them to their landing pages and build a research dashboard while you do other work. It reads video ads by exporting a frame for every second and transcribing the audio.