How to Find and Analyse Competitor Ads in 2026 (Meta Ad Library, TikTok Center, and AI Spy)

Competitor ads are the highest-leverage research input in performance marketing. Most Superscale customers running paid spend at scale treat competitor-ad analysis as a recurring weekly workflow, not an annual audit. The 2026 stack to do this well combines five free public ad libraries (Meta, TikTok, LinkedIn, Google, X), one purpose-built scraping layer, and a structured framework for turning what you find into testable variants in your own account.

This guide is the operator's playbook for that workflow. It covers where to find competitor ads on every major platform, the eight signals to extract from each ad you find, how to convert competitor research into a test matrix, the legal and ethical edge cases that come up, and the workflow Superscale customers like SumUp, marketbirds, and Advercy use to compress competitive research from "we'll look at it next quarter" to "we look at it every Monday morning."

This post pairs with our piece on static ads for the creative patterns that recur in competitor research, and our agentic marketing definition for where competitor analysis fits in the autonomous loop.

What we mean by "competitor ads"

A competitor ad, in the strict sense, is any paid creative being actively or recently run by a brand that competes with you on the same audience, the same channel, the same objective, or some combination. The looser sense, and the one most operators actually need, extends to any paid creative from any brand whose creative strategy can teach you something. A DTC mattress brand can learn from an apparel brand on hook patterns. A B2B SaaS company can learn from a consumer fintech on social-proof formats. The competitor frame is more about creative-pattern relevance than industry membership.

Concretely, the things people mean when they say "find competitor ads" usually include:

Ads currently running on Meta, TikTok, LinkedIn, Google, X, Pinterest, Snap, or Reddit.
Recently-paused ads still visible in the platform ad library.
Top-spending ads by run-length proxy (longer-running ads imply better creative performance).
New creative concepts from a tracked competitor list.
Trending formats inside a vertical or theme.
Ad copy, hook patterns, CTAs, and offer structures.

Each of those is a distinct query, and 2026's tooling landscape solves them differently. The remainder of this post walks through the answers.

Where to find competitor ads in 2026: the full ad-library stack

The five platform-native ad libraries below are public, free, and required reading for any operator doing this work. The third-party tools on top of them are optional, but most teams above $50k/month spend pay for at least one. The treemap below shows where most operator research time gets spent.

Where operators spend competitor-research time — Superscale 2026 customer survey, n=84

┌──────────────────────────────────────────────────────────────────────────┐
│ Meta Ad Library (facebook.com/ads/library)            ████████████  38%  │
├──────────────────────────────────────────────────────────────────────────┤
│ TikTok Creative Center (ads.tiktok.com/business/...)  ████████      24%  │
├──────────────────────────────────────────────────────────────────────────┤
│ Third-party ad spy tools (Foreplay, Atria, AdSpy)     █████         15%  │
├──────────────────────────────────────────────────────────────────────────┤
│ LinkedIn Ad Library (linkedin.com/ad-library)         ███            9%  │
├──────────────────────────────────────────────────────────────────────────┤
│ Google Ads Transparency Center (adstransparency...)   ██             7%  │
├──────────────────────────────────────────────────────────────────────────┤
│ AI-driven competitor analysis (in-platform agents)    ██             5%  │
├──────────────────────────────────────────────────────────────────────────┤
│ X (formerly Twitter) Ad Repository                    █              2%  │
└──────────────────────────────────────────────────────────────────────────┘

1. Meta Ad Library

URL: https://www.facebook.com/ads/library.

Coverage: every active ad on Facebook, Instagram, Messenger, and the Audience Network globally. Political and issue ads are kept for seven years; commercial ads disappear when the advertiser stops running them.

Capabilities:

Filter by country (required), platform (FB, IG, Messenger, Audience Network), language, advertiser name, keyword, date range, and media type (image, video, carousel, no creative).
View the ad creative, ad copy, CTA, and the destination URL.
See start date (the first time the ad ran), the single most useful proxy for ad performance.
See variations of the same ad when the advertiser is using dynamic creative.

Best for: depth on Meta paid spenders. The library is the most complete on the open web for any platform; if a brand is running paid Meta, this is where you find it.

Limitations: no spend data for commercial ads, no impression counts, no CTR. You are inferring performance from run length and variation count.

2. TikTok Creative Center

URL: https://ads.tiktok.com/business/creativecenter/.

Coverage: top-performing TikTok ads globally, surfaced by TikTok's own performance algorithm. Coverage is curated, not exhaustive. You only see what TikTok decides to show.

Capabilities:

Filter by industry, region, ad objective, ad format, like count, CTR rank, conversion rank, and date range.
View the ad creative (video plays in-line), engagement stats (likes, comments, shares), and a directional CTR rank.
See the audio used (linked to TikTok's commercial sound library), the duration, and the call-to-action button.
Side-by-side compare two ads.

Best for: trend discovery, hook patterns, format research. Volume is moderate but signal density is high because every ad you see has been pre-filtered for performance.

Limitations: no impression counts, no spend data, no ability to see ads from a specific competitor unless they happen to be in the top performers.

3. LinkedIn Ad Library

URL: https://www.linkedin.com/ad-library/.

Coverage: ads currently or recently active across LinkedIn placements (Sponsored Content, Message Ads, Conversation Ads, Document Ads, Lead-Gen Forms).

Capabilities:

Search by advertiser, keyword, country, and date.
View the ad creative, headline, intro copy, and CTA.
See the advertiser's company page.

Best for: B2B competitor research. Coverage is thinner than Meta because LinkedIn's ad volume is smaller, but for B2B SaaS, agency, and enterprise plays, this is the only place that surfaces single-image sponsored content.

Limitations: launched in mid-2023, still has lower discoverability than Meta's library, and search returns less consistently sorted results.

4. Google Ads Transparency Center

URL: https://adstransparency.google.com/.

Coverage: ads on Google Search, YouTube, Google Display Network, Discovery, Gmail, and Shopping, surfaced per advertiser globally.

Capabilities:

Search by advertiser name or domain.
Filter by ad format (text, image, video, shopping).
View ad creatives, see geographic distribution, and see the most recent ad date.

Best for: Google Ads competitor research. Particularly useful for understanding which Google formats a competitor is investing in (Search vs. Display vs. YouTube vs. Performance Max).

Limitations: no spend data, less granular date filtering than Meta, and YouTube video ads can be hit-or-miss to play in-line.

5. X (Twitter) Ad Repository

URL: https://ads.x.com/repository.

Coverage: ads currently or recently running on X.

Capabilities:

Search by advertiser, region, keyword, date.
View the ad creative and copy.

Best for: niche use cases. X's ad ecosystem in 2026 remains a fraction of Meta/TikTok/LinkedIn for most B2C and B2B brands.

Limitations: thin coverage, intermittent uptime, no spend data.

Third-party ad spy tools

Above the platform-native libraries, three categories of paid tooling exist in 2026:

Ad spy databases. Foreplay ($79–$249/month), Atria ($99–$399/month), AdSpy ($149/month), BigSpy ($99–$249/month). These scrape and index public ad libraries plus their own panel data, offer Boolean search, save-to-folder workflows, and integrations with Slack/Notion.
Creative analytics suites. Motion, Triple Whale Creative Cockpit, Northbeam Creative Analytics. These pull your own and competitors' ad data alongside your performance metrics. Less about discovery, more about analysis.
AI-driven competitor agents. The newer category. Superscale's built-in Competitor Tool, Vexpower's competitor briefs, AdCreative's competitor research module. These do the discovery, scoring, and pattern-extraction in one step, then feed the output into the next-asset generation step.

The choice of tier depends on volume. A solo founder running a single brand can do this work for free using the platform-native libraries plus a manual Notion tracker. A growth team at $100k/month spend gets serious leverage from a third-party tool. An agency managing 5+ client brands gets compounding returns from an AI-driven agent that handles the discovery-to-asset flow end-to-end.

What to extract from a competitor ad: the eight signals

Looking at an ad is not the same as analysing one. The eight signals below are what experienced media buyers extract from every competitor ad they find. Skip these and you'll waste hours scrolling through ad libraries with no testable output.

1. The hook (first 1.5 seconds for video, headline for static).

Write down the literal text. If the hook names a specific friction or a specific number, flag it. Those are the patterns that compound. Vague hooks ("Discover better marketing") are not patterns worth replicating.

2. The visual subject.

What does the eye land on first? A face, a product, a screenshot, a chart, a hand holding an object. Write it down in two words.

3. The format.

Static, video, carousel, dynamic creative, collection ad, catalog ad. Each format buys different things. (See our static ads guide for the format-by-format breakdown.)

4. The aspect ratio.

9:16 vs 4:5 vs 1:1 vs 16:9. Patterns vary by ratio. A 9:16 hook is different from a 4:5 hook.

5. The CTA.

The exact wording of the call to action. "Shop Now," "Learn More," "Get Free Trial," "Apply Now," "Download." Each implies a different funnel stage.

6. The offer or value proposition.

What is being promised. Price, time savings, status, identity, transformation. If you can't summarise the offer in one sentence, the ad is unclear and probably not high-performing.

7. Run duration.

How long the ad has been active. In Meta Ad Library, check the "started running" date. Anything past 60 days is likely a winner being scaled. Sub-7-day ads are tests; weight them less.

8. Variation density.

How many versions of this ad exist (same creative, different copy; same copy, different visual; same everything, different audience). High variation density implies serious investment in this concept.

A template extraction worksheet most teams use looks like this:

Field	Example: Taxfix DE Meta static
Advertiser	Taxfix GmbH
Platform	Meta (Facebook Feed)
Format	Static, 4:5
Hook	"Steuererklärung in 22 Minuten — wie diese Berliner-Mutter dabei 1.200 € zurückbekommen hat"
Visual subject	Forum-thread style screenshot
CTA	"Jetzt starten" (Get started)
Offer	"22 minutes" + "€1,200 average return"
Run duration	84 days
Variation density	7 visible variants
Pattern label	Discussion-thread / specific-number-hook
Hypothesis	The discussion-thread layout out-performs poster-style on German Meta for tax audiences

That last row, the hypothesis, is what turns a research exercise into a test. Without it, you have a folder of screenshots. With it, you have a brief.

Converting research into a test matrix

The bridge from "I have looked at competitor ads" to "I have shipped a new winning creative" is a structured test matrix. The simplest one we use looks like this:

Test matrix for static ad iteration — example, three-week sprint

┌─────────────────────────────────────────────────────────────────────────┐
│ Pattern source    │ Hook variant      │ Visual subject  │ Test week    │
├───────────────────┼───────────────────┼─────────────────┼──────────────┤
│ Discussion-thread │ Specific number   │ Forum screenshot │ Week 1      │
│ Discussion-thread │ Question-led      │ Forum screenshot │ Week 1      │
│ Street-interview  │ Specific number   │ Interview frame  │ Week 2      │
│ Street-interview  │ Identity-led      │ Interview frame  │ Week 2      │
│ Apparent screenshot│ Surprise         │ Phone screen     │ Week 3      │
│ Apparent screenshot│ Curiosity gap    │ Phone screen     │ Week 3      │
└─────────────────────────────────────────────────────────────────────────┘

Each cell becomes one creative. Six creatives, two patterns per week, three weeks of testing. By the end of week 3 you have CPA data on six concepts, you know which pattern won, and you can scale the winners while shipping six new variations.

Two rules that compound this matrix:

Vary one variable at a time. Within a pattern, keep visual + format constant and change only the hook. Across patterns, keep hook variable constant and change only the visual frame. Without this discipline, you can't attribute the win.
Re-test winners across markets. A pattern that wins for one geo or audience often wins for adjacent ones, but it has to be re-tested before you can scale. Taxfix's UK street-interview pattern moved to DE only after a re-test confirmed the pattern held.

The Superscale customer workflow

Across customer accounts at scale, a recurring weekly workflow emerges:

Monday: scan. 30 minutes in Meta Ad Library, TikTok Creative Center, and the in-platform Competitor Tool. Mark 10–15 new ads for analysis.
Monday: extract. Fill the 8-signal worksheet for each marked ad. 5–7 minutes per ad. Output: 10–15 hypothesis rows.
Tuesday: brief. Turn the top 3 hypotheses into briefs. One sentence each: pattern, hook variable, visual subject, format, ratio.
Tuesday: generate. Feed briefs into the creative system (designer plus Figma, or an agentic generator). Output: 6–9 statics ready for review.
Wednesday: review and launch. Human review for brand alignment, then push to the ad platform. Launch budget at $200–$500/day per creative.
Friday: read. Pull CPA from the platform. Statistical readout against the prior week's winners. Update tracker with the new patterns.
Following Monday: repeat.

SumUp's content team, running 120+ Meta ads across 8 languages by the time of their published case study, uses this loop with minor modifications. marketbirds' agency operation runs it across five client brands in parallel. Advercy, the one-person consultancy, runs it for five clients alone, using Superscale to compress the brief-to-generate-to-publish steps into roughly 15 minutes.

The compression matters. A team using Meta Ad Library plus Figma plus manual Ads Manager upload spends 6–8 hours on this loop. A team using an integrated agent like Superscale's Competitor Tool plus the publish-back path compresses it to around 90 minutes. The number of times per month you can run the loop is the rate at which your creative learning compounds.

How we define "high-signal competitor ad" (methodology disclosure)

Not every ad you find in a library is worth analysing. Treating them all as equal noise is the most common mistake newer operators make. A high-signal competitor ad meets four criteria:

Run duration ≥ 30 days. Anything shorter is likely a test, not a scale ad. Performance signal is weak.
Variation density ≥ 3 visible variants. A single creative implies low investment. Three or more variants implies the advertiser is iterating on this pattern.
Same advertiser has multiple high-duration ads. A brand running one long-duration ad is a fluke. A brand running five long-duration ads is a system.
Pattern is replicable in your own brand voice. A street-interview pattern works for a fintech because the audience is general consumer. The same pattern fails for an enterprise B2B database because the audience reads, not watches.

Roughly 15–20% of ads in any platform library meet all four criteria. The other 80% are noise: tests that didn't work, brand-awareness pieces that aren't direct-response, or one-off campaign creatives. Filtering aggressively saves hours.

Legal and ethical edges

Three questions come up repeatedly when teams systematise competitor-ad research:

Is scraping competitor ads legal?

Public ad libraries (Meta, TikTok, LinkedIn, Google, X) are explicitly public. Viewing, screenshotting, and writing down what you see is unambiguously fine. Automated scraping at scale enters a grey zone. The platforms' terms of service generally restrict it, but the case law on enforcement is thin. Third-party ad spy tools handle this through scraping plus public-API access; you outsource the legal exposure to the tool when you use one.

Can I use a competitor's exact ad copy?

No. Ad copy is copyrightable. Translating a pattern (the discussion-thread layout, the specific-number hook) is fine. Copying literal sentences is not. The fastest way to get a brand-safety complaint or a takedown is to lift verbatim text from a competitor.

What about competitor product imagery?

Hard no. A competitor's product photography is their copyright. Using it in your ad is both a copyright violation and a trademark violation if their packaging is visible. Compose your own product imagery, even when the competitor pattern is what you're studying.

A defensible workflow: study the pattern, document the structural elements, write your own hook, shoot or generate your own visual. Everything you ship is original creative inspired by, not copied from.

Edge cases the libraries miss

Five categories of competitor activity that platform-native libraries don't capture well:

Email marketing. Klaviyo, Customer.io, and Mailchimp campaigns are private. You see them only by subscribing to the competitor's list. Worth doing: sign up to your top 10 competitors' lists from a dedicated inbox.
Push notifications. App-side. Capture by installing the competitor app and enabling notifications.
SMS campaigns. Phone-side. Same approach as push.
Affiliate and partnership creative. Creative being run by affiliates, influencers, and partners under the brand's program. Visible by searching for the brand handle on TikTok, Instagram, and YouTube, but not in the brand's own ad library.
Programmatic display below brand-named domains. A brand running display through DV360 or The Trade Desk often appears as a different advertiser name in transparency reports. Requires triangulation.

For most brands, the first three are where the additional signal lives. Subscribe, install, and check weekly.

What changes in the next 18 months

Two shifts to plan around:

Library throttling. Meta has signalled tighter rate limits on the Ad Library API and tighter geographic filters on the public web interface. Expect more friction by Q1 2027. Workflows that depend on heavy programmatic scraping should plan a fallback to manual review or licensed third-party data.
AI-driven competitor briefs replace manual extraction. Agents that ingest a competitor's ad-library output and produce a one-paragraph brief ("this brand is testing discussion-thread layouts with specific-euro-amount hooks; you should test the same pattern with US-dollar amounts and an English-language audience") are crossing into reliability. The 5–7 minute manual extraction step becomes a 30-second prompt.

How to start, this week

Five steps to go from "we don't track competitor ads" to "we have a weekly cadence" in under five workdays:

Make a list of 10 competitors. Direct competitors plus 3 brands you admire creatively in adjacent verticals.
Set up the libraries. Bookmark Meta Ad Library, TikTok Creative Center, LinkedIn Ad Library, Google Ads Transparency Center. Save 10 search URLs (one per competitor) for one-click access.
Build the extraction template. A Notion table or a Google Sheet with the 8 signal columns. Pre-fill with 5 example rows from a competitor you know well.
Run the first loop. One hour, Monday morning. Extract 10 ads. Produce 3 briefs. Hand to creative (or feed into an agentic generator).
Track the new creative against the briefs. Did the discussion-thread pattern hold for you? Did the street-interview pattern translate? Iterate.

If you want the loop pre-built (auto-discovery of competitor ads, structured pattern extraction, direct hand-off to creative generation, publish-back to Meta, TikTok, and Instagram) that's what Superscale's Competitor Tool plus the broader agent was built to do. Free credits on sign-up; first competitor brief in under 5 minutes.

Frequently asked questions

What is the best free tool to find competitor ads?

Meta Ad Library, by a wide margin, for any brand running Meta. TikTok Creative Center for trend research. LinkedIn Ad Library for B2B. Google Ads Transparency Center for Search/YouTube/Display. All four are free, public, and indexed by advertiser.

How do I find every ad a specific competitor is running?

Search the advertiser's exact name in Meta Ad Library with all ad formats enabled and no date filter. For TikTok, search the brand handle in TikTok Creative Center and in TikTok's main search; some ads only show in one. For LinkedIn, search by company page. For Google, search the brand domain in Google Ads Transparency Center.

Can I see how much a competitor is spending on ads?

Not from public sources, for commercial ads. Meta discloses spend only for political and issue ads. SimilarWeb, SensorTower, and AppFigures estimate via panel data with wide error bars. Third-party ad-spy tools sometimes show paid-spend estimates but none are reliable to better than ±50%.

Is there a TikTok Ad Library?

Yes: TikTok Creative Center (ads.tiktok.com/business/creativecenter/). It is curated to top performers, not exhaustive. For exhaustive TikTok coverage, third-party tools like Foreplay and Atria index more broadly.

What about competitor ads on Instagram specifically?

Instagram ads are included in Meta Ad Library. Filter by "Instagram" under the Platforms filter. Coverage is identical to Facebook ad coverage because they share the same ad system.

How often should I check competitor ads?

Weekly is the cadence we see in customer accounts that compound. Monthly works for slow-moving B2B categories. Daily is excessive for almost everyone and creates more noise than signal.

Should I copy competitor ad patterns directly?

Patterns, yes. Copy, no. A discussion-thread layout is a pattern in the public domain. The exact sentences inside it are copyrightable. Study the structure, write your own words.

What's the difference between competitor research and ad intelligence?

Competitor research is a manual process focused on a tracked list of brands. Ad intelligence is the broader category of structured creative-data analysis: your own ads, competitor ads, top-performing patterns in your vertical, scored against performance benchmarks. Our piece on digital ad intelligence covers the latter.