Studio Handbook
This page explains when to use normal generation, batch generation, and workflow mode, plus the purpose of each workflow node in English.
Modes
Each Studio mode supports a different level of speed, scale, and process control. Start with the simplest option that fits the job.
Use this mode when you want to create one result at a time and adjust the prompt, model, or settings interactively.
Best for: Quick experiments, prompt tuning, one-off visuals, video drafts, and audio generation with manual control.
Use batch mode when you need many outputs in one run using a shared setup with multiple prompts or prompt-image pairs.
Best for: Campaign variations, product sets, A/B creative testing, marketplace assets, and large prompt lists.
Use Workflow when you want reusable pipelines made of connected nodes instead of repeating the same steps manually.
Best for: Team processes, repeatable production flows, multi-step prompt refinement, and structured generation pipelines.
Use Cases
These are the most typical reasons teams use Studio in production and day-to-day creative work.
Generate one concept quickly, compare models, and refine prompts in short feedback loops.
Produce many banner, product, or social variations with batch mode and consistent parameters.
Turn repeatable creative operations into reusable node-based pipelines for consistent output quality.
Use uploaded images, multi-image references, brand style analysis, or edit-region workflows for tighter control.
Workflow Nodes
Use this as a quick reference when building flows. Each node has a specific role in the pipeline and should be chosen based on the job to be done.
Purpose: Adds one reference image to the workflow.
When to use: Use it when a generation or editing step needs a single image input.
Purpose: Adds one reference video file to the workflow.
When to use: Use it when later nodes need source footage or a motion reference.
Purpose: Adds one reference audio file to the workflow.
When to use: Use it for voice, sound, or music-based downstream steps.
Purpose: Collects multiple reference images in one node.
When to use: Use it when style, composition, or product details come from several images at once.
Purpose: Extracts a brand style profile from one or more reference images.
When to use: Use it when you want more consistent visual language across many generated outputs.
Purpose: Stores the main instruction text for image, video, or audio generation.
When to use: Use it whenever you need a direct text prompt in the workflow.
Purpose: Stores supporting text such as USP, product notes, story context, or background information.
When to use: Use it when extra context should travel through the workflow without being the main prompt.
Purpose: Refines, restructures, and standardizes a long prompt before generation.
When to use: Use it when raw prompts are messy, inconsistent, or need brand style merged in.
Purpose: Creates a generation-ready prompt from images and a short brief using an LLM.
When to use: Use it when users provide rough input and need the system to write the final prompt.
Purpose: Produces images from prompt input and optional references.
When to use: Use it for the main image generation step in a visual pipeline.
Purpose: Produces video from prompt input and optional image or frame references.
When to use: Use it when the workflow needs animated or cinematic output.
Purpose: Produces speech, music, or other audio outputs from prompt and voice settings.
When to use: Use it for TTS, voice workflows, or prompt-based audio generation.
Purpose: Improves image resolution for larger delivery sizes.
When to use: Use it after generation when the result needs sharper export quality.
Purpose: Resizes or crops an image to a target frame.
When to use: Use it when outputs must match a required aspect ratio or placement size.
Purpose: Separates the product or subject from the background.
When to use: Use it for ecommerce assets, clean cutouts, or compositing workflows.
Purpose: Edits a selected area of an image with inpainting.
When to use: Use it when only part of the image should change instead of regenerating the whole frame.
Purpose: Combines multiple looks or generated images into a collection preview.
When to use: Use it when reviewing a series, campaign set, or storyboard sequence together.
