What the Meta Ad Library doesn't show you (2026)

What the Meta Ad Library does show

Start with what works, because it is a lot. Meta's help page says the library includes all currently active ads running on Meta products, and it is open to anyone: no account is needed to search it, and it costs nothing. You can search by advertiser or by keyword, and every ad you open shows its creative, its text and the platforms it runs on.

That is why it is the first stop for competitor ads research. If you want to know what a brand is running this week, it is the one complete and current source, because it comes from Meta itself. The rest of this guide is about what it leaves out.

Everything below was read on 6 November 2026 from Meta's own transparency, help and developer pages, listed under Sources.

1. Spend and reach exist only for political and EU ads

This is the limit people hit first. Meta's Ad Library tools page says spend, reach and the funding entities are shown for ads about social issues, elections or politics. For everything else, a shoe brand, a supplement, a SaaS app, there is no spend figure and no reach figure.

The developer reference for the ArchivedAd object makes the split exact. These fields come back only for social issue, election and political ads: spend, impressions, currency, bylines (who funded the ad), estimated_audience_size, demographic_distribution and delivery_by_region. Both spend and impressions are reported in ranges, not exact numbers.

Ads delivered in the EU get a different set. The reference lists eu_total_reach and beneficiary_payers for EU ads only, and the age, gender and location breakdowns for UK and EU ads. So an ecommerce ad that happened to reach someone in Germany shows reach and targeting detail, while the same ad shown only in the US shows none.

What this means for you: you cannot use the library to work out what a competitor spends. Anyone quoting a spend figure for an ordinary ad is estimating, not reading it from Meta.

2. Ads disappear once they stop running

The second gap matters most if you build a swipe file. Meta shows every ad while it runs. Once an ad stops, only social issue, election and political ads (kept 7 years) and ads delivered in the EU (kept a year after their last impression) stay in the library.

Meta's help page says the same from the other side: only issue, election and political ads remain searchable after they stop running. Its Ad Library tools page gives the figures: seven years for social issue, election and political ads whether active or inactive, and one year after the last impression for ads delivered in the EU.

Meta's pages that we read give no retention period for ordinary commercial ads shown outside the EU. What they do say is that the library covers active ads for those, so the safe working assumption is that a stopped US-only ad is gone. Plan your research around that: capture what you want while it runs.

A related catch: the API's ad_active_status filter defaults to ACTIVE, so even programmatic searches return only live ads unless you ask for inactive or all.

3. There is no performance data

The library tells you what is running. It does not tell you what is working. There is no click-through rate, no conversion count, no cost per result, no return on ad spend, for any ad. The API field list has none of them either.

That is why a long run is the best public signal and still only a signal. Brands rarely keep paying for ads that lose, so an ad that has run for months deserves a closer look. But a brand can leave a weak ad on by accident, run an awareness ad that is judged on other things, or keep a test going. Treat run length as a reason to study an ad, never as proof that it made money.

4. The start date and the version count are hints, not facts about results

Each ad in the library carries a start date. The API field behind it, ad_delivery_start_time, is documented as the time the advertiser wants delivery to begin, and its ad_delivery_stop_time is blank when the ad runs until it is stopped or the budget is spent. A separate field, ad_creation_time, is described by Meta as not the same as when the ad ran.

Two practical consequences:

• A start date is not a performance date. An ad that "started" in March and still shows is not necessarily one that ran steadily. An advertiser can pause and resume, and the card does not tell you.
• The date is the ad's, not the concept's. A brand that re-uploads the same creative as a new ad restarts the clock. A long-running concept can hide behind several short-dated copies.

Many ads also carry several versions of the same creative: the API returns lists for the text on "each unique ad card" (ad_creative_bodies, ad_creative_link_titles). On the website, an ad with more than one version says so. More versions suggests the brand is testing variations, which is worth noting in your log. It is not a count of how many tests won.

5. Search is built for auditing, not for browsing

The library is a transparency tool first. You can search a keyword or an advertiser, and filter by country, platform, media type and active status; the API takes the same filters plus language and delivery dates. Three things follow from how it is built.

• Keyword search matches the whole ad. Meta documents that it searches text, images, audio from video and call-to-action buttons. That is why a keyword returns other advertisers who merely use the word.
• Searches are not translated. Meta's reference says to write keywords in the language of the ads you want. A Spanish-language ad will not show for an English query.
• There is no niche, format or hook filter. You cannot ask for "UGC skincare ads" or "ads with a testimonial hook". You find a brand first, then read its ads.

6. Links die and nothing can be saved or followed

The ad's address works while the ad runs. When the ad stops, the address leads to nothing for anything outside the retention rules above. A bookmark, a Notion link or a spreadsheet row that points at a stopped ad is a dead link, and the creative you meant to study is gone with it.

The library has no way to save an ad: no save button, no folders, no boards. It does have Saved searches: on keyword results and advertiser pages, "Save search" keeps the query and your filters under a name, so you can rerun it. That is a bookmark, not a follow. There are no alerts that we found. Nothing tells you a competitor launched a new ad, so the check is a habit you build yourself.

These three gaps together are the reason people look for Meta Ad Library alternatives. The routine in the next section covers each of them for free, by hand.

7. The funnel and the targeting are out of sight

For ordinary ads the library shows the ad, not what happens after the click. You see the creative and the call-to-action button, but not the landing page experience, a quiz in front of it, or a pre-sell page. To see them you click through yourself, from a clean browser, and note what you find.

Targeting is also thin. The ArchivedAd reference lists target ages, target gender and target locations for UK and EU ads, and for political and issue ads delivered to Brazil. For other ads there are no interests, no lookalikes and no audience definition. You learn the ad, not who it was aimed at.

How to work around the limits for free

You can cover most of these gaps with a spreadsheet and thirty minutes a week. For the full routine with exact filters, see how to track a competitor's new Facebook ads every week. The short version:

1. Pick three to five competitors and find each one's Page in the library. Bookmark the results with the country, active-status and platform filters set.
2. Check once a week, on the same day. Look at the start dates for anything new since last time.
3. Capture before it stops. Screenshot the image, screen record the video, and paste the primary text, headline and button into your sheet along with the Library ID. This is the only way to keep a non-EU ad (how to save ads from the Facebook Ad Library).
4. Log run length yourself. Write the date you first saw the ad and the date you last saw it. That gives you a real run length that the library will not keep.
5. Click through once. Open the landing page and note the offer and the page type.
6. Say what you do not know. In the sheet, keep a column for "why we think it works" and write it as a guess, because there is no result data to confirm it.

This works and costs nothing. It also takes time that grows with every brand, and an ad that starts and stops between two visits is missed.

What keeps ads after Meta removes them

If the gaps above cost you time, the options split into three kinds. You can see all of them side by side in Meta Ad Library alternatives and the best ad swipe file tools.

• Tools with a browser extension save an ad from the library into your own collection in one click, so you keep it yourself. See Foreplay and Swipekit.
• Ad spy tools run their own crawlers and keep a history, usually with estimated metrics. They are the most expensive kind, and their metrics are the tool's estimate, not Meta's data.
• A library that is already saved. This is AdSwipeFile: winning Facebook and Instagram ads filed by brand, niche and format, with each creative kept on our own servers after Meta takes the ad down. You do not collect anything. It currently holds 63,244+ ads.

Where AdSwipeFile and your AI agent fit

AdSwipeFile does not remove Meta's limits, and it does not claim to. It works around three of them.

• Ads that stay. A library ad keeps its creative after Meta removes it. Browse by niche: Beauty, Fashion, Fitness and wellness, Food and drink, Pets, Home and garden, Apps and software, Jewelry and watches. By format: UGC selfie, Review, Before and after, Product image, Unboxing, Listicle, Text overlay, Us vs them. By brand: AG1, Glossier, Liquid Death, Warby Parker, or every brand. Ads from followed brands that we track on Meta also keep a copy of the creative.
• Follow and save. Paid plans follow 10 brands on Pro ($39 a month) and 50 on Agency ($99), with a 7-day trial. Anyone with a free account can heart ads to save them; shareable boards come with Pro and its trial.
• Your AI agent does the searching. Connect AdSwipeFile to Claude, ChatGPT or Cursor and ask "show me the longest-running UGC ads in pet food and write three hooks from them". Free accounts get 50 MCP/API calls for the life of the account.

What it does not fix, plainly: it shows no spend, reach or results (the library holds none, and neither does Meta), it covers Facebook and Instagram only, it has no Chrome extension, and it holds far fewer ads than the big spy tools. For a brand it has not filed, the Meta Ad Library is still where you look first.

Try AdSwipeFile with your assistant

Copy the setup prompt and paste it into Claude, ChatGPT or Cursor. It connects the library and asks for recent ads from your niche. Free: 50 MCP/API calls for the life of the account and the first 50 ads of every list, no card.

Copy setup promptCopied

Continue with Google

Setup steps for each assistant

FAQ

Does the Meta Ad Library show how much advertisers spend?

Only for ads about social issues, elections or politics, where spend and impressions come as ranges. For ordinary commercial ads there is no spend figure, and any number you see elsewhere is an estimate.

Can I see a competitor's old Facebook ads after they stop running?

Not in the Meta Ad Library, except for social issue, election and political ads (kept 7 years) and ads delivered in the EU (kept a year after their last impression). To keep other ads, capture them while they run or use a tool that stores them.

Does the Meta Ad Library show ad performance like CTR or conversions?

No. It has no click-through rate, conversions or return on ad spend for any ad. How long an ad has run is the nearest public signal, and it is a hint, not proof of results.

What are the best Facebook Ad Library alternatives that keep ads after they stop running?

Tools with a Chrome extension such as Foreplay or Swipekit save ads into your own collection. AdSwipeFile files winning ads by brand, niche and format and keeps each creative after Meta removes the ad. See the compare page for the full list.

Is there a follow button or alert in the Meta Ad Library?

Not that we found. You search the advertiser again each time. A tool that follows brands, or a weekly routine with a spreadsheet, fills that gap.

Can my AI agent search the Meta Ad Library for me?

Meta has no MCP server that we found, and its API is limited to political, issue and EU ads. An AI agent connected to AdSwipeFile searches its own library of Facebook and Instagram ads instead.

Is the Meta Ad Library still worth using?

Yes, for what a specific advertiser runs right now, free and straight from Meta. Use it for current checks and pair it with a saved library for history.

Sources

Checked on 6 November 2026 on each tool's own site. Prices are in US dollars as listed, and change: check the source before you buy. AdSwipeFile's facts come from its pricing page.

1. Meta Ad Library
2. Meta Ad Library API
3. Meta Ad Library tools
4. Graph API ads_archive reference
5. Foreplay pricing
6. Foreplay MCP
7. Foreplay API
8. Foreplay home page
9. Foreplay mobile app
10. Swipekit pricing
11. Swipekit ad library MCP
12. Swipekit home page
13. Meta Ad Library tools (transparency.meta.com)
14. Meta Advertising Standards: Ad Library
15. What is the Meta Ad Library and how do I search it? (Meta Help Center)
16. Graph API reference: ArchivedAd fields