Launched this week
SCRIPTAU
The easiest way to turn your voice into short-form scripts
30 followers
The easiest way to turn your voice into short-form scripts
30 followers
Hit record. Talk about your video idea. SCRIPTAU writes a script that sounds like you, not like AI. No prompts, no templates. Your voice in, your script out.
really digging the focus on retention. does the ai look for specific 'viral patterns' when structuring the script, or is it just following a standard video format?
@mika_guedes Great question! It's definitely not a standard video format, we built a pattern library from analyzing viral content across different genres (MrBeast, branded content, challenge videos, comedy confessions, social transformation stories, etc.) and extracted the actual structural mechanics that drive retention.
The chat feature after the script is generated is interesting. Can I use it to ask for a completely different version of the same audio, like turning a long story into a 15-second teaser, or is it just for small tweaks?
@marinacarneiro The refinement feature is designed for surgical tweaks, adjusting a specific scene, changing the tone of a line, or reworking a section. You talk to it like a director giving notes: "make the opening more aggressive" or "cut the middle section."
For a completely different format like turning a long story into a 15-second teaser, you'd get better results just
recording a new take and saying something like "15-second teaser version of your idea", the engine will build it
from scratch with the right pacing for that length.
Loved the app! Already created 6 scripts and they look amazing!
Most tools try to make scripts sound professional, but for tiktok, that's a death sentence. How does this handle the 'hooks'? I need something that grabs attention in the first 2 seconds, not a slow intro!
@mysty_flow So glad you're loving it! 6 scripts already, that's awesome.
And you're 100% right, "professional" is the enemy of retention on TikTok. We built the hook system specifically around that.
Every script starts with what we call a visual-behavioral hook, not a voiceover, not a title card, not "Hey
guys." It's an image or action that creates tension before a single word is spoken. The AI is trained on patterns
we extracted from analyzing viral content across genres, and the #1 rule is: the first beat has to make someone
stop scrolling with their thumb, not their ears.
So instead of "In this video I'm going to show you..." you get something like "close-up, hand trembling holding the envelope, silence" a visual moment that makes the viewer ask "what's happening?" before anyone speaks. And the WHY IT RETAINS column in each script actually shows you the retention mechanic behind every single beat, so you can see exactly why the hook works and learn the pattern for future content.
TL;DR: the AI treats the first 2 seconds as the most important part of the entire script. If the hook doesn't stop a thumb, nothing else matters.
the fact that there's no sign-up is such a breath of fresh air for a 6-day build. i'm curious about the 'brain dump' logic though. is there a limit to how long the audio can be for it to still generate a coherent script, or can i just talk for 5+ minutes and it will summarize the best parts?
@harley_machado Thanks! And yeah, no sign-up wall, we wanted people to just talk and see the magic before deciding anything.
On the brain dump question: right now on the free tier you get 60 seconds of audio, which is plenty for a quick idea. Pro gives you 5 minutes, and Pro+ goes up to 15 minutes, so yes, you can absolutely ramble for a while.
But here's the key part: it doesn't just summarize. It restructures. If you talk for 5 minutes jumping between
ideas, repeating yourself, going on tangents, the AI extracts the core narrative, finds the emotional beats, and
rebuilds it as a timed, scene-by-scene script with camera direction and retention logic baked in.
So it's less "here's a summary of what you said" and more "here's the script a director would write after listening to your pitch." The longer you talk, the more raw material it has to work with.
@henri_sarcinella the app uses your device's native speech recognition, which handles normal background noise pretty well (coffee shop, street, etc). but for best results, a quieter spot definitely helps, especially with longer recordings. the key thing is: you don't need to speak perfectly. even fragmented ideas get transformed into
a full script.
love the clean ui. i noticed you mentioned high-retention scripts, does the refinement logic suggest different hooks or structures based on the platform (like tiktok vs. youtube shorts), or is it more about fixing the tone of the voice-over itself?
@josemartins____ right now the refinement focuses on the script itself, tone, structure, pacing, hooks. it doesn't auto-switch formats based on platform (tiktok vs youtube shorts), but you can tell it exactly what you want: 'make this more tiktok', 'add a stronger hook', 'make it shorter for reels', and it adjusts. platform-aware presets are on the roadmap though