You point AdWhispr at a competitor. It ingests their whole Meta ad library, tracks it day over day, and hands you a read on what they are actually running right now: which ads have survived the longest, what hooks and formats they lean on, and where the gaps are. It never guesses spend or invents a CTR. The one honest signal it trusts is ad longevity, because an ad a brand keeps paying to run is an ad that is working.
Use this when
You are about to compete against a specific brand, entering a category with an entrenched player, or refreshing your own creative and want a factual read on the market instead of a guess from their website copy. It is also the right first move before any of the cloning or launch skills, because everything downstream is better when it starts from a proven ad rather than a blank page.
Do not use this to judge your own account performance. This skill only reads public competitor signals. For your own numbers, use the Ad Account Audit skill instead.
What you need
- The competitor's brand or Facebook page name as it appears in the Meta Ad Library. Legal entity names often differ from the consumer brand, so have both ready if you know them.
- AdWhispr connected to your AI client. If it is not connected yet, the box at the top of this page shows the one-time setup for Claude, ChatGPT, Cursor, and the rest. No ad account of your own is required for a competitor read.
- A few minutes for a single brand. The first ingestion of a large advertiser takes longer while the full library is pulled and classified.
How it runs
- Confirm the target with the user: the exact brand or page name, and the market if it matters, since the ad library is region aware.
- Resolve the brand and ingest its full Meta ad library, then classify each ad for hook, format, tone, and offer.
- Rank the ads by days running. The longest-running ads sit at the top, because sustained spend is the most honest proxy for what is working.
- Read the shape of the account: total active ads, the split across static, carousel, and video, and how the format mix has shifted over the tracked window.
- Categorize the angles across the winners, so the output is a pattern (price, urgency, social proof, product led, lifestyle) rather than a list of copy.
- Hand back a structured teardown, with every claim traceable to the ad library, and a short note on the openings the competitor is leaving on the table.
What the skill checks
- Every claim about an ad (active status, first-seen date, copy, format) is read from the tracked library, never assumed from an earlier session, because a competitor can pause or launch mid-week.
- No spend figure is stated anywhere. Ad count, format mix, and longevity are the only quantities this skill has, and none of them convert reliably to a dollar amount.
- The market filter is confirmed before concluding a brand is "not advertising," since an empty result with the wrong region selected is a false negative, not a finding.
When it stops
- Done: the teardown covers the requested brand with longevity-ranked winners, format mix, angle categories, and gap notes.
- Nothing to report: the brand has no active ads in the requested market right now, which is itself a real signal and is reported as such with the date.
- Blocked: the brand name does not resolve to any advertiser after reasonable variants are tried. The skill asks you for the exact page name or URL rather than guessing.
What it will and won't do
Read only. This skill never touches an ad account, never spends, and never contacts anyone. It reads public ad-library data and returns a brief. Any next step you want to take from it, cloning a winner or launching against it, is a separate skill with its own approval.
What you get
A structured competitive brief: the competitor's longest-running ads with days-running attached, the format and platform emphasis, the angle and offer patterns across their winners, how their volume has trended over time, and an honest list of the angles or formats they are not running that you could. No spend estimates, no invented performance scores, no verdict beyond what the data supports.
FAQ
How do you know which competitor ads are working without spend data?
By how long each ad has been running. Brands pause ads that lose money and keep the ones that pay, so an ad still live after months is the closest honest signal to a winner. AdWhispr tracks every ad's first- and last-seen dates daily and ranks by that, rather than inventing a CTR or ROAS it cannot see.
Where does the ad data come from?
The public Meta Ad Library, ingested per brand and tracked day over day. Nothing here requires access to the competitor's account, which no tool has.
Can it analyze more than one competitor at once?
Yes. Run it per brand, then use the Head-to-Head Brand Comparison skill to put two competitors side by side.
Does this touch my ad account?
No. It is read only and competitor facing. Your own account is never queried by this skill.