Make a Video From a Script You Already Have — Story Mode

You already have a script

📌 The screens in this post are UI mockups built from a real project's data and the app's own CSS. The values in them (scenes, segments, prompts) are the actual output. The project itself ran in Korean; the story text shown here is an English translation of that output — the numbers, timings, and prompts are unchanged.

Whether you wrote it yourself, commissioned it, or drafted and polished it in another tool — if the script is finished, there's no reason to ask Story mode to write one from scratch. Paste the script, and Story mode takes it as-is, has you confirm the cast, and continues from scene splitting.

In the stepper above, only 1 Research is grayed out (disabled) — with a script in place there's no topic to research. 2 Synopsis stays open, because that's where the cast is extracted back out of your pasted script and confirmed. The flow is 3 Script (done) → 2 confirm the cast → 4 Scene split.

Where to paste the script: the Setup tab

At the bottom of the 0 Setup tab there's a Script (optional) field. Just paste your script there. Pick the same options — title, output language, scene-split unit — but the moment you fill the script field, the flow switches from "write a new one from a title" to "run with this script."

• Scene-split unit — decide whether to split finely per sentence or group by time (seconds). Choose to match your script length and editing style.
• Output language — match it to the language of your script (Korean or English).
• Title — worth filling in to keep the project organized.

Research is skipped; Synopsis just confirms the cast

In a project that already has a script, the Research gate tab locks — the topic is settled, so the step is moot.

Synopsis does not lock. Its job changes: instead of deriving a storyline, it extracts the cast back out of your pasted script and asks you to confirm it. You can fix each character's name, gender, age, and appearance here, and those values flow straight into voice casting in step 5 and the character descriptions in step 6. Until you confirm, ▶ Run all stays disabled — a guard against running the downstream steps on an unsettled cast.

From scene split to audio and prompts

From here it's identical to the title-only flow. The values in the screens below are the output of a real project, "The Rich Man and the Poor Man," where one script became 11 scenes, 20 segments, 72.5 seconds.

4 Scene split — the pasted script becomes a table

Rows are segments, not scenes. A scene with several segments repeats its number in the # column.

• Speaker — lines you prefixed in the script, like Gatekeeper:, are picked up as speakers. Marking dialogue before pasting is what makes them split apart here.
• Emotion — the table shows an emotion only on non-narrator dialogue ((Angry), (Sad)). Narration has an emotion value too, but it isn't displayed and is always synthesized as normal — only dialogue emotion reaches the TTS in step 5.
• SFX rows — sound-effect cues arrive as their own tinted rows. They come out as English descriptions like heavy iron gate slamming shut because the sound-effect model takes English.
• Split again — if the result is too fine or too clumped, change the unit and min–max seconds in the bar above and split again. Your script stays untouched; only the table is rebuilt.

5 Audio — a voice per speaker

• Voice picker — browse Typecast, Gemini, and ElevenLabs voices in one window.
• Per-speaker progress — a counter like 16/16 shows how many of that speaker's segments are done. You can also generate one speaker first and listen.
• Segment status — every row carries a done badge, and Test regenerates just that one line so you can hear it.
• Your own recordings work too — if you already have takes, drop an mp3 and an SRT onto a speaker row and that file is used instead of TTS. It pairs well with bringing your own script.

6 Prompts — an image and video prompt per scene

Each scene gets one image prompt and one video prompt. Note the @Park-style mentions inside them.

Turn on [auto] for Scene split, Audio, and Prompts, press ▶ Run all, and the rest runs on its own.

Character @mentions still work

Even in a pasted script, characters register themselves as reference cards. Scenes featuring them get the mention inserted automatically. The card's image slot starts empty, so generate the reference image once and it is attached to every mentioned scene from then on, which is what keeps characters consistent. You can add or remove them by typing @ in a prompt; names with a space are wrapped in braces, like @{Chairman Yun}.

Generate and export

Once the prompts are ready, generate images and videos and export to CapCut, Adobe Premiere, or Vrew in one click — timeline, subtitles, audio tracks, and Ken Burns animations all included.

In short

If you have a finished script, there's no need to rewrite it. Setup tab → paste the script → confirm the cast → Run all → export. The two starting paths meet at exactly the same point after scene splitting.

Want to see one episode end to end with its real output? Read From one title to 11 scenes — a real run. No script yet? See how to start from a single title; want to nail the topic first? See YouTube benchmarking.