The problem this solves
TikTok performance lives or dies in the first two seconds, and the knowledge of what hooks are currently working in your category is sitting in the public ad library - in video form, which is exactly why nobody extracts it. Studying ten competitor ads properly means watching each repeatedly, transcribing the hooks, and decoding why the opening holds attention. That is hours of focused work the creative team never gets between production deadlines.
Without it, TikTok creative gets brainstormed from general best practices and whatever went viral last month somewhere else. The category-specific intelligence - which hook patterns your actual competitors are sustaining spend behind, which emotional registers and formats are working against your shared audience - never enters the room. Teams ship guesses at a platform that punishes guessing faster than any other, and creative iterations burn budget rediscovering what the library already showed.
How the mission runs
- Search the TikTok ad library. Ad Libraries sweeps TikTok's public ad archive for your category and surfaces the creatives with the strongest sustained runs - the ads whose longevity signals they keep earning their budget against your shared audience.
- Transcribe the hooks. Video Transcripts extracts what each ad actually says and shows in its opening seconds: the spoken hook, the on-screen text, the visual device. The part of each ad that does the heavy lifting becomes analyzable text.
- Decode the angle patterns. The agent classifies the ten ads by hook mechanism, emotional register, format, and offer framing - and identifies the patterns: which openings recur across advertisers, which are one brand's signature, and which plausible angles nobody in the category is using.
- Post the breakdown to #creative. The full analysis lands in your #creative Slack channel: each ad with its transcribed hook and classification, the pattern summary, and the open angles worth testing. The creative team gets the intelligence where they already work, ready to react to.
The prompt
This is the exact objective the agent receives. Swap the obvious placeholders for your own domain, segment or channel and run it as-is from the console, Slack, or the API.
What comes back
An angle breakdown posted to your #creative Slack channel: the ten strongest-running TikTok ads in your category, each with its transcribed hook and a classification of the mechanism behind it, plus a synthesis of which patterns dominate, which are saturated, and which angles remain open. Your next TikTok production cycle starts from the category's live evidence instead of a blank brief.
Make it yours
- Widen the pull: analyze twenty-five ads for a full quarterly creative-intelligence review rather than a production-cycle briefing.
- Track one rival: run the same breakdown on a single competitor's TikTok presence to decode their entire hook playbook.
- Make it a rhythm: rerun monthly and have each post highlight what changed - new hooks entering, old ones retiring.
Frequently asked questions
How is "strongest-running" determined?
By sustained presence in the library. Public archives show what runs rather than conversion data, so longevity under continued spend is the honest proxy for performance - advertisers rarely keep paying for openings that stopped working. The breakdown is explicit that this is inference.
Should we imitate the winning hooks directly?
Use the mechanism and change the surface. The breakdown identifies why each hook works - curiosity gap, pattern interrupt, blunt claim - so your team can run the proven mechanism through your own voice and offer. Direct imitation of a saturated hook mostly buys blindness in the feed.
Does this connect to actually making the creatives?
It is the intelligence half of the loop. The breakdown arms your team or a follow-up mission to script and produce counter-creative from the identified angles - and companion playbooks in this category pick up exactly there, generating clips from the hook analysis.