A swipe file you never open is just a folder
Everyone in paid media saves competitor ads. Almost nobody uses them. The difference is not the tool — it is that a screenshot records what an ad looked like and loses the only three things that make it useful later.
What a screenshot throws away
You see a competitor's ad running for the fourth week and think: that is working.
You screenshot it. Six weeks later you go looking and you have a cropped JPG named
Screenshot 2026-04-11 at 14.32.png that tells you nothing.
Three things were lost.
The original asset. A screenshot of a video is a frame. The pacing, the first three seconds, the captions — the parts that were actually doing the work — are gone. Same with a carousel: you captured card one of five.
The destination. The ad was half the funnel. What the landing page promised, how the form was structured, what the offer really was — none of it is in the screenshot, and by the time you look the page has changed.
The context. When you saw it, how long it had been running, which platform, whose it was. Without that, you cannot tell a tested winner from something launched on Tuesday.
Structure it so a stranger could use it
The test is whether someone who did not build the file could find the right example in under a minute. That requires three decisions, made once.
Name files predictably. Date, brand, platform, format, sequence.
07_22_26_LUMEN_STUDIO_01 sorts chronologically, groups by brand and survives
being dragged into a shared drive. Anything relying on your memory will fail.
Store the ad and the page together. One folder per capture, containing the creative and a full-page PDF of where it pointed. This is the single highest-value habit in the whole practice, and it is the one everybody skips because capturing a full page used to be a chore.
Add one line of note. Not analysis — an observation. "Running 5+ weeks, offer is financing not price." Thirty seconds now saves the whole capture from becoming decoration.
What to actually look for
A swipe file is not a mood board. You are collecting evidence about four things, and it helps to know which one you are looking at.
- Offers, not aesthetics. What is being promised, and at what commitment level? Free inspection versus fixed price versus financing. The offer decides more than the design does, and it is the easiest thing to test yourself.
- Longevity. An ad running for weeks has survived optimisation. One launched three days ago has survived nothing. If the platform shows you a start date, that date is the most useful field on the screen.
- The first three seconds. For video, the hook is the whole game. Collect hooks separately from full ads — you will reuse them far more often.
- Funnel length. Does the click go to a form, a call, a quiz, a calculator? That tells you what commitment level they found works in your market, which is expensive information to buy yourself.
The friction problem, and how we removed it
All of the above is well known and almost nobody does it, for one boring reason: right-click does not download a Meta video, carousels take five separate captures, and a full landing page needs stitched screenshots. Ten minutes per ad. So people take one cropped screenshot and move on.
We got tired of it and built the tool, then gave it away. AdVault is a free Chrome extension that downloads the original creative — video, image or whole carousel — from Meta, TikTok and LinkedIn with clean filenames, and captures an entire landing page as PNG, JPG or PDF with auto-scroll. It runs locally in your browser: no account, no tokens, nothing passing through a server of ours. About five seconds per ad instead of ten minutes.
Being direct about the incentive: we built it because we needed it, it is genuinely free with no plans or limits, and there is nothing inside it to upsell. If the habit above is worth having, the friction should not be what stops you.