How does the Competitor Ad Intel tool work?

Enter a competitor's website URL. The tool fetches their active ads from the Meta Ad Library API (the official public database of all Meta ads), captures ad screenshots, and runs vision analysis on each screenshot to extract creative style, hook type, funnel stage, production polish, and stealable elements. You see real ad data plus descriptive vision labels.

What is real data vs vision-derived labels?

Real data: ad count, ad copy, ad images, run cadence, format mix, target locations and ages. All from the Meta Ad Library API. Vision-derived labels: creative style, hook type, hook technique, emotional tone, target audience, funnel stage, production polish, creative notes, stealable elements. Vision labels describe what is in the ad; they do NOT claim the ad performs well or works.

Does this tool estimate competitor ad spend?

No. The Meta Ad Library does not expose impression counts, click counts, conversions, or spend. The tool returns ad count, age distribution, format mix, and creative-pattern frequencies. It does not estimate, forecast, or guess spend, revenue, ROAS, or any performance figure.

How is this different from Meta Ad Library directly?

Meta Ad Library shows raw ads with no analysis. The tool adds vision-based creative analysis on a sample of the active ad set: hook type, creative style, funnel stage, stealable elements, and a structured playbook. Vision is the labeling layer; Meta is the data layer.

Can the tool analyze any brand?

Any brand running active Meta (Facebook/Instagram) ads, since the Ad Library is publicly searchable. Brands not running Meta ads return zero results. Currently Meta only; Google, TikTok, and other platforms are out of scope.

What are stealable elements?

Stealable elements are descriptive observations of creative components visible in the analyzed ads: hook structures, overlay styles, CTA placements, format choices. They are suggestions for what another brand could test in their own A/B slate. They are not guaranteed performance predictions.

How fresh is the data?

Each scan is a real-time pull from Meta Ad Library; no caching across scans. Meta typically reflects new ads within 24-48 hours of go-live. Vision analysis runs at scan time on up to 8 fresh screenshots.

IndexSee which brands AI assistants actually recommend.Who AI actually recommends.See all

Read a competitor's Meta ads (TikTok coverage coming).

Pulls every active ad a competitor runs on Meta from the public Ad Library and uses vision analysis to extract creative style, hook, funnel stage, and stealable elements.

MayaMemory