# Best Websites to Find High-Quality Ad Swipe Files > The best swipe file websites for DTC, TikTok, B2B and search ads, what each one hides, and three filters that separate proven ads from ads you merely liked. - **Author**: Marxx - **Published**: 2026-10-24 - **Category**: Competitive intel - **URL**: https://marxx.ai/posts/best-swipe-file-websites --- **What are the best websites for high-quality ad swipe files?** It depends on what you're making. Use the Meta Ad Library for DTC social ads, TikTok's Top Ads for TikTok-native video, the LinkedIn Ad Library for B2B, and Google's Ads Transparency Center for search. Each source has a bias, so filter for proof: ads that lasted, were copied, and fit your buyer. Then test, don't copy. ## There's no single best swipe file website A good sommelier doesn't recommend "the best wine". They ask what you're eating, then pick the wine that goes with it. Swipe file websites work the same way. A DTC video ad, a B2B LinkedIn ad and a Google search ad need different sources. But there's a second thing most "best swipe file sites" lists miss. **Every source shows you a filtered slice of reality, and each one filters differently.** One shows only ads that are still running. Another shows only ads advertisers chose to display. Another shows ads that someone found impressive. If you don't know a source's filter, you'll draw the wrong lesson from it. So this guide gives you: 1. A pairing menu: what you're making, and where to look. 2. The bias in each source, and how to correct for it. 3. Three filters that separate high-quality swipes from ads you merely liked. 4. A worked example from search to test. A pairing menu: DTC video ad with Meta Ad Library, TikTok-native ad with TikTok Creative Center, B2B ad with LinkedIn Ad Library, search ad with Google Transparency Center, anything fast with the Marxx library, copy inspiration with swipe-file newsletters --- ## Part 1: The pairing menu The pairing menu: what you're making, where to look and what to look for | You're making... | Look in | Cost | What to look for | |---|---|---|---| | **A DTC video ad** | Meta Ad Library | Free | The oldest still-running ads, especially with many near-identical versions | | **A TikTok-native ad** | TikTok Creative Center, Top Ads | Free | The first two seconds, and where engagement peaks on the per-second graph | | **A B2B ad** | LinkedIn Ad Library | Free | The offer: a report, a template, a demo | | **A search ad** | Google Ads Transparency Center | Free | The claims every competitor repeats, so you can say something else | | **Anything, fast** | Marxx library | Paid | Long-running ads in your category, tagged by hook, offer and format | | **Copy inspiration** | Swipe newsletters and archives | Mostly free | Structure, not wording | --- ## Part 2: Know what each source hides A swipe source is a sample. And every sample has a selection rule: something that decides which ads you see and which you don't. Knowing the rule tells you what you can conclude. Every swipe source has a bias: Meta shows survivors, TikTok Top Ads is curated, LinkedIn shows recent ads, Google is partial, newsletters reflect taste, Marxx shows signals ### Meta Ad Library: Survivors **What it shows:** Every ad currently running on Facebook and Instagram, with its start date. **Its bias:** For ordinary commercial ads shown outside the EU, it keeps ads **only while they're active**. When an ad stops, it disappears. Ads delivered in the EU are kept for a year after they last ran, with targeting and EU reach. Political and issue ads are kept for seven years ([PPC Land](https://ppc.land/ads-library/)). The same library, two views: outside the EU only live ads with no reach or targeting; inside the EU ads kept for a year with targeting and reach **What that means:** Outside the EU you see what's running *now*, never what was tried and abandoned. That's survivorship (below). **How to correct for it:** Sort by start date and look for ads that have run a long time. For a wider view of what a brand tested, look at its EU-served ads, which stay visible after they stop. ### TikTok Top Ads: Curated **What it shows:** A collection of high-performing creatives, filterable by region, industry and campaign objective. Each ad has a per-second engagement graph showing where viewers engaged most ([TikTok](https://ads.tiktok.com/help/article/top-ads?lang=en)). **Its bias:** TikTok says all featured ads are **authorised by advertisers** inside TikTok Ads Manager. So you see the ads brands chose to show, not a random sample of what's running. **How to correct for it:** Use it for **format and pacing**, such as how the first two seconds work and where attention peaks. Be cautious about concluding which *messages* work best in your category. ### LinkedIn Ad Library: Recent, with detail mainly for the EU **What it shows:** Ads from roughly the last year, searchable by advertiser, keyword, country and date range. Ads served in the EU show impressions by country and detailed targeting; outside the EU, details are limited ([Superads](https://www.superads.ai/blog/linkedin-ads-library)). **Its bias:** About a year of history, with no performance or spend data. **How to correct for it:** Look at **offers** and **audiences**. In B2B, the offer (a report, a template, a demo) often matters more than the design. ### Google Ads Transparency Center: Partial **What it shows:** Ads from verified advertisers, searchable by advertiser, region and format, with the date each ad was last shown ([Google, 2023](https://blog.google/technology/ads/announcing-the-launch-of-the-new-ads-transparency-center/)). **Its bias:** It tells you what ran, not what worked. **How to correct for it:** Use it to map the **claims** your category repeats in search, then find the one nobody makes. ### Newsletters and galleries: Taste **What they show:** Ads someone found impressive. **Their bias:** Editors pick ads that look clever, which isn't the same as ads that sold. **How to correct for it:** Borrow **structure**, such as how an argument is built, not conclusions about what works. ### The trap in every "top ads" list: Survivorship bias During the Second World War, statistician Abraham Wald pointed out that the aircraft returning with bullet holes showed where a plane *could* be hit and survive. The planes hit elsewhere never came back. That's **survivorship bias**: drawing conclusions only from what made it through. Survivorship bias: in a grid of ads you only see the few survivors; long-running ads tell you what a brand kept, not what the same idea did for everyone who tried it **In swipe files:** A long-running ad tells you what *one brand kept*. It doesn't tell you how the same idea performed for everyone who tried it and stopped. That's why the right response to a great swipe is **test, don't copy**. > **The quotable version:** A swipe file shows you what survived. It can't show you what died trying the same thing. That's why you test, not copy. --- ## Part 3: What makes a swipe "high quality" Most swipe files are full of ads someone liked. A high-quality swipe is one with **evidence** behind it. Use three filters before you save anything. Three filters before you save: it lasted, it was copied, it fits you 1. **It lasted.** Still running after 30 days or more. Brands switch off ads that don't work, so a long run means the brand kept paying for it. In the Marxx library, Wonderla's "Buy 3, Get 1 Free" ads had been live for **99 days** in October 2026. 2. **It was copied.** The brand runs several near-identical versions, a sign it's scaling or testing around a winner. Those Wonderla ads had **35 to 44 copies each**. 3. **It fits you.** Same kind of buyer, offer or format as the ad you're about to make. A theme park's offer stack can inspire a software ad, but only if you translate the *structure* to your buyer. **Where these filters mislead:** A long run can also mean an evergreen awareness ad, or one nobody remembered to switch off. Many copies can just mean the same ad split across regions. That's why the third filter, fit, and your own test matter. Real long-running competitor ads with days live and copies, October 2026 An ad that passes all three filters is worth more than a hundred you just liked. --- ## Part 4: Worked example *Hypothetical example: a B2B software team needs a LinkedIn ad.* Worked example: pair with LinkedIn Ad Library then Meta, filter for 30+ days and the same buyer, check EU-served ads for targeting, extract the offer, and test report-led against demo-led 1. **Pair:** Start in the LinkedIn Ad Library, then check Meta for B2B brands that also run there. 2. **Filter:** Keep ads from brands selling to the same buyer that have run for 30+ days. 3. **Correct for bias:** Look at EU-served versions to see the targeting detail other regions hide. 4. **Extract the pattern:** Several long-running ads offer a benchmark report built around one striking number. 5. **Test:** A report-led ad against your usual demo-led ad, with the same budget and audience. **What you've learned either way:** If the report wins, you've found an offer your buyer responds to. If it loses, you've learned the category pattern doesn't transfer to your buyer. That's the survivorship trap caught by a test, not a quarter of wasted spend. --- ## Part 5: Save it so you can find it again Finding the ad is half the job. The other half is finding it again next month. The Marxx browser extension saves any ad in one click and tags it by hook, offer, emotion and format, with days live. Later you can search for "price hook, UGC, still live" instead of scrolling a folder. Our [swipe file tools guide](/blog/swipe-file-tools) explains why most swipe files become graveyards, and how to fix yours. Marxx creative board grouping saved ads by hook, angle and format Marxx trending tactics showing which hooks are rising in a category --- ## Six mistakes, and how to spot them | Mistake | Symptom | Fix | |---|---|---| | One source for everything | B2B ideas from DTC galleries | Pair the source with what you're making | | Ignoring a source's bias | Copying curated "top ads" as if typical | Know each source's selection rule | | Saving ads you like | A swipe file full of taste, not evidence | Filter: lasted, copied, fits | | Treating long-running as proven | Copying an evergreen awareness ad | Combine days live with copies and fit | | Copying instead of testing | Me-too ads that flop | Test the structure against your current ad | | Copying words | Ads that sound like a competitor | Borrow structure, never wording or branding | ## What to take into your next creative review - **Pair the source with what you're making.** DTC, TikTok, B2B and search each have a best library. - **Every source has a selection rule.** Survivors, curated, recent, partial or taste: know which you're looking at. - **Save evidence, not taste.** Lasted, copied, fits. - **A swipe shows what survived, not what died trying.** Test, don't copy. - **Borrow structure, not words.** ## FAQ **What's the best website for ad swipe files?** It depends on what you're making. The Meta Ad Library for social ads, TikTok's Top Ads for TikTok, the LinkedIn Ad Library for B2B, and Google's Ads Transparency Center for search. **Are there free swipe file websites?** Yes. All the official ad libraries above are free. **How do I know if an ad is worth saving?** Check that it has run for 30+ days, that the brand runs several versions, and that it fits your buyer and offer. **Can I copy a competitor's ad?** Borrow the structure, like a price-first hook or an offer stack. Don't copy the words, images, logo or brand. **Why do some ads disappear from the Meta Ad Library?** Outside the EU, Meta keeps ordinary commercial ads in the library only while they're running. --- ## Save any ad, auto-classified Install the Marxx extension: one click saves any ad from any of these sites, tagged by hook, offer, emotion and format, with how long it has run. **[Install the Marxx Extension: Save any ad, auto-classified ->](https://www.marxx.ai)** *Related: [Tools and Extensions for Managing Ad Swipe Files](/blog/swipe-file-tools) - [Free Ad Spy Tools](/blog/free-ad-spy-tools) - [What Features Should a Reliable Ad Spy Tool Have?](/blog/ad-spy-tool-checklist)* ### Sources - PPC Land. [Meta Ad Library: What it shows and how long ads are kept](https://ppc.land/ads-library/). - TikTok. [Top Ads](https://ads.tiktok.com/help/article/top-ads?lang=en). - Superads. [LinkedIn Ads Library guide](https://www.superads.ai/blog/linkedin-ads-library). - Google (2023). [Announcing the Ads Transparency Center](https://blog.google/technology/ads/announcing-the-launch-of-the-new-ads-transparency-center/). - Marxx ad library, October 2026. --- - [More Competitive intel articles](https://marxx.ai/posts/category/competitive-intel) - [All articles](https://marxx.ai/posts)