Script in, scenes out. Paste your voice-over (up to about 15 minutes of speech) and get it cut into timed scenes that fit your video model’s clip length, each with its narration and a standalone video prompt you can paste straight into Veo, Kling or Sora.
A worked example
Script (opening): “Deep in the Arctic, a lonely lighthouse keeper hears a knock at midnight. Nobody has visited the island in eleven years. He lifts his lantern, opens the door, and finds… a polar bear cub, shivering in the snow.”
Settings: Clip length 8 s (Veo) · Model Veo 3 · Aspect 16:9 · Main character: a 70-year-old fisherman with a grey beard, yellow raincoat and blue wool hat.
- Scene 1 · 0:00–0:08 · Narration: “Deep in the Arctic, a lonely lighthouse keeper hears a knock at midnight.” · Prompt: wide cinematic shot of a lone lighthouse on a snowy rock at night, aurora above, beam sweeping the sea; inside, the keeper (70-year-old fisherman, grey beard, yellow raincoat, blue wool hat) looks up from a mug of tea at a knock…
- Scene 2 · 0:08–0:16 · Narration: “Nobody has visited the island in eleven years. He lifts his lantern…” · Prompt: close-up of the same keeper raising a brass lantern, warm light on his face, snow swirling past the window…
- Scene 3 · 0:16–0:24 · Narration: “…and finds a polar bear cub, shivering in the snow.” · Prompt: the door opens to a tiny polar bear cub on the step, blue moonlight, breath visible…
Each scene repeats the character description, so the keeper looks the same in every clip.
How to use it
- Paste the narration. One paragraph or sentence per idea works best. The tool estimates speaking time from the words.
- Choose the clip length your model makes: 8 seconds for Veo, 10 seconds for Kling and Sora, or 5, 6, 15 or 20 seconds for others.
- Choose the model, aspect ratio and visual style, and describe your main character once (optional). It is repeated in every clip.
- Split instantly (free) to get timed scenes with simple prompts, made in your browser. Or Write scene prompts with AI (one free run) for cinematic, model-specific prompts per scene.
- Generate each clip in your video model, then join them in any editor, or send the script to Script to Video Studio for a picture-and-voice version.
Who it is for
- Documentary and storytelling channels working from a written script: script prompts and YouTube prompts.
- Faceless channel owners turning narration into consistent visuals: faceless channel prompts.
- Teachers and explainer makers: explainer prompts and prompts for teachers.
What it does not do
- It does not render video. It plans the clips and writes the prompts.
- Timings are estimates from the word count. If you record your own voice-over, adjust scenes to your real pace (Script to Video Studio does this automatically from your recording).
- It will not fix a weak script. If a scene has nothing to show, rewrite that line to describe something visible.
- Video models still drift between clips. Keep the character, lighting and style words identical and regenerate any clip that breaks continuity.
Privacy
“Split instantly” runs in your browser and sends nothing. “Write scene prompts with AI” sends your script and settings to Google Gemini for that one request; nothing is saved on PromptifyLab.
Tools that go before and after
- Viral Video Prompt Builder or Video to Prompt: get the idea or storyboard first.
- Script to Video Studio: turn the same script into a finished video with pictures, voice and subtitles.
- YouTube Title & Hook Generator: publish it with a strong title and hook.
Questions
How long can my script be?
Up to about 15 minutes of speech. Longer scripts work best split into parts.
Which clip length should I choose?
Match your model: 8 seconds for Google Veo, 10 seconds for Kling and Sora 2, 5 or 6 seconds for many others. Shorter clips give each scene one clear action.
Is it free?
Splitting is free and unlimited. Writing AI scene prompts uses one free daily AI run, or unlimited runs with your own Gemini key.
Can I use it for languages other than English?
Yes. The narration stays in your language; the video prompts are written in English because video models follow English best.