Launching today
Scrollsnacks
Launch videos, broken down second by second
20 followers
Launch videos, broken down second by second
20 followers
Launch videos broken into timed hook, problem, solution, feature, proof and CTA sections. Search them by brand or section, or from your AI tools via MCP.
Free Options
Launch Team
Dropbox Sign APISimplify your eSignature integration
Promoted
Hi Product Hunt, I'm Lorenzo, and I built Scrollsnacks!
Whenever I needed a launch video, I scrubbed through other launch videos to see how long the hook runs, when the product shows up and whether there is an ask at the end. I also wanted my AI to read those references before drafting a script. So I built a library we can both use.
What it does:
Real launch videos, each split into timed sections like Hook, Problem, Feature, Proof and CTA
Plug it into Claude, Cursor or Codex with the Scrollsnacks MCP, so your agent can study and reuse what works when it drafts a script
Browse free by brand or category; Pro adds section filters and natural language search
Every video credits the brand, and the studio when we can confirm it
One thing surprised me while building it: about half of all launch videos never actually ask you to do anything at the end. No "try it", no link, nothing. Easy win if you're making yours.
It's free to start. Full breakdowns are part of Pro. Tick the product news box at sign-up before 15 Nov for 7 days of Pro, no card needed.
Feedback welcome: what would you ask the library first, and what's missing?
https://scrollsnacks.com/
This would be really useful when putting together a launch video. Sometimes the hardest part is just figuring out what to include and how to structure everything. Congratulations!
congrats on the launch lorenzo. to your question, first thing i would ask is which hooks actually held people past three seconds, not only which ones looked nice. if the library also showed how each cta performed that would make it useful for more than editing )
@agshin_rajabov Thanks, that’s the right next question.
Today we show how hooks and CTAs are built (timing, techniques, on-screen text), not whether they held attention or converted. We don’t get brand-side retention or click data from the source films.
We only ship films that already proved reach on the source platform, clear a craft bar (real launch work, at minimum 720p, no AI slop), and pass a two-person section review before they go live.
True “held past 3s” / CTA performance is still on the wishlist if makers share outcomes.