What are the 5 steps in the ScriptMotion workflow?

The 5 steps are Format & Layout, Character & Product, Background & Environment, Concept & Script, and Audio. Step 1 locks in an AI model based on your layout. Steps 2 and 3 configure visuals — the AI presenter, product placement, and the generated background. Step 4 captures your script with scene breakdowns. Step 5 collects voice, delivery style, and background music preferences in a single pass — passed as hints to the selected video model, which may honor them.

Which AI video model is used at each step?

The model is selected once, at step 1, based on the layout you pick. Hyper-Realistic uses Seedance 2. Product Marketing and Multi-Shot both use Kling 3.0 Omni Pro. Custom lets you pick any of the thirteen models directly, including Sora 2 Pro for realistic UGC-style footage. Steps 2 through 5 then configure inputs that feed into the single selected model when you generate.

Can I skip steps or jump back to edit earlier choices?

The workflow is structured but not strictly linear — you can navigate back to any earlier step to change a choice. Skipping ahead is constrained because later steps depend on the model and format selected at step 1. Going back to step 1 and changing the layout will re-lock the model, which may invalidate format options like aspect ratio, resolution, or duration.

How long does a full 5-step run take end-to-end?

Filling out all five steps typically takes 10 to 20 minutes depending on script length and how much customisation you do. Generation itself runs asynchronously through Trigger.dev and most clips finish in 1 to 3 minutes for short ads, longer for 4K Veo renders. You can leave the page while a generation runs — progress is delivered via server-sent events.

What happens if a generation fails mid-workflow?

Generations run on Trigger.dev with retries built in. If a specific AI provider experiences issues, the circuit breaker trips and notifies you immediately rather than silently switching to a mismatched model. You can re-run the generation, adjust inputs, or switch to the Custom layout to pick a different model. Credits are only consumed when a generation produces a usable output.

Can I save a workflow and reuse it for the next campaign?

You can save your progress mid-build with Ctrl+S at any step, so a draft you start today can be picked up tomorrow. Reusing the same configuration across multiple campaigns is straightforward — duplicate the draft and adjust the script or product for the new campaign. A dedicated template library is not yet shipped; save-and-resume is the primary persistence mechanism today.

5 steps. Every video. Every time.

The guided workflow builder turns chaotic creative requests into a repeatable video factory. Each step enforces platform constraints and tracks completion. Save your progress with Ctrl+S.

Format & Layout

Choose a layout style — Hyper-Realistic (powered by Seedance 2), Product Marketing or Multi-Shot (both powered by Kling 3.0 Omni Pro), or Custom (pick any AI model directly, including Sora 2 Pro for realistic UGC-style footage). Each layout locks in a dedicated AI model and constrains format options to what that model supports best.

Once you select a layout, aspect ratio, resolution, and duration options automatically adjust to the assigned model's capabilities. The Custom layout exposes all 10 AI models for power users who want full control.

Character & Product

Choose from a global library of AI presenters or generate a custom character. Add product placement with compositing or 3D rotating display styles.

Custom character generation uses AI image generation (5 credits). Select products from your brand's product library and configure placement position and animation.

Background & Environment

AI-generated backgrounds with custom prompts or preset environments. Configure color grading, camera angles, and scene composition.

Choose AI-generated or preset environments and set lighting, color treatment, and style references for the overall visual aesthetic. (Brand-kit enforcement — color palette and voice-and-tone — is applied later, at step 4.)

Concept & Script

Write or paste your script, then break it into scenes with speakers and duration hints. Live word count and estimated duration keep you on track.

Each scene can have a dedicated speaker, text cues, and timing hints. The system estimates total video duration based on speech rate and scene count.

Audio

Set voice and background music preferences in a single step — source, language, accent, tone, pacing, and a music description with a narration/music volume balance. These are passed as hints to the selected video model, which may honor them in the generated output.

Per-speaker voice settings with pacing (Slow, Normal, Fast), emphasis words, pause markers, and a pronunciation guide for brand terms, plus a music description (mood and genre) — all provided as prompt guidance to the chosen video model. Whether spoken audio or music appears in the final clip depends on the model's audio capability.

Frequently asked questions

Ready to streamline your video production?

Start Creating