Writing prompts for Sora 2 is harder than creating video itself.. Here is a workaround :-)
One thing I’ve noticed after playing with Sora: writing good *video prompts* is way tougher than the video generation itself.
With images, even a half-baked idea can give something usable. But with video, if you don’t clearly define *scene, style, and camera work*, the results can feel random or chaotic.
The pain points I kept running into:
* Too vague → jittery, incoherent clips
* Overstuffed → Sora ignores half the details
* No structure → I’d rewrite the same prompt 10 different ways
The turning point was realizing video prompts need a *template-like structure* (scene + style + camera + movement). Once I started approaching it this way, my outputs felt way more cinematic and consistent.
For example, instead of writing “a car chase,” a structured prompt becomes:
>
Much better results.
I’ve been experimenting with this free tool above that helps break down prompts into those sections.
Has anyone else found a good framework for writing video prompts?
tipseason•1h ago
One thing I’ve noticed after playing with Sora: writing good *video prompts* is way tougher than the video generation itself.
With images, even a half-baked idea can give something usable. But with video, if you don’t clearly define *scene, style, and camera work*, the results can feel random or chaotic.
The pain points I kept running into:
* Too vague → jittery, incoherent clips * Overstuffed → Sora ignores half the details * No structure → I’d rewrite the same prompt 10 different ways
The turning point was realizing video prompts need a *template-like structure* (scene + style + camera + movement). Once I started approaching it this way, my outputs felt way more cinematic and consistent.
For example, instead of writing “a car chase,” a structured prompt becomes:
>
Much better results.
I’ve been experimenting with this free tool above that helps break down prompts into those sections.
Has anyone else found a good framework for writing video prompts?