Adult fashion direction
Joi AI SIJoi AI SI
An ai photo editor turns a single reference still into a directed visual rewrite, keeping the structure you need while reimagining the finish. VideoAny puts Seedream 5 Pro behind a straightforward image-to-image workbench where you upload a source, describe the changes, and review the result. Every generation uses account credits, with the task cost shown before submission, so you can plan a batch of catalog refreshes, room restages, or packaging updates without guesswork. This page walks through the controls, the creative decisions that matter, and the inspection steps that turn a good result into the right one.
The model you select shapes how faithfully the output follows your reference and how freely it interprets your description. Seedream 5 Pro is the selected image model for this page, and it rewards clear keep-and-change instructions with a result that feels connected to the source without being a simple filter pass. The three controls below set the foundation before you write a single word of prompt.
01
Seedream 5 Pro is the selected image model for this page, and it works best when you treat the reference as a structural guide rather than a pixel map. An ai photo editor should let you describe what stays and what transforms, and this model reads both signals together. Choose it when the job needs a believable material finish, consistent lighting, and a composition that holds its ground while the details change. An ai photo editor that skips model selection leaves too much to chance, so confirm Seedream 5 Pro is active before you upload.
02
The image-to-image workflow accepts a reference image and a description of the desired visual changes, but the aspect ratio you set determines whether the output fits its destination. An ai photo editor can stretch or crop a result that was composed for a different frame, so match the ratio to the final use before you submit. If the reference is a square product shot and the output needs to fill a 4:5 campaign board, set that ratio early. An ai photo editor that lets you change the ratio after generation forces a second round of inspection.
03
Completed tasks appear in generation history for preview and download, giving you a side-by-side moment to inspect framing, subject detail, and visual consistency. An ai photo editor does not read your intention; it reads your prompt and your reference, and the gap between them shows up in the output. If a material looks plastic when it should feel matte, adjust the description and run again. An ai photo editor rewards iteration, and the workbench keeps your previous settings so you can refine one clause at a time.
The workbench gives you a reference upload, a prompt field, and model and aspect-ratio selectors. Together they form a tight loop: choose a source that already holds the structure, write what must stay and what is allowed to change, and restyle a draft without starting from a blank page. Each control is explained below with a concrete visual subject so you can see how the decision flows into the result.
01
A neutral living room with sofa, plants, and even daylight makes a strong reference because the spatial layout is clear and the lighting is predictable. An ai photo editor reads the arrangement of objects, not just the colors, so a cluttered source confuses the model. Pick a reference where the composition already works and the surfaces are legible. This online photo editor workflow depends on a clean starting point, and a well-chosen reference cuts the number of retries in half. An ai photo editor can reinterpret materials and palette, but it cannot rescue a source where the structure is unreadable.
Keep the living room layout, sofa, plants, and camera height from the uploaded photo. Change only the lighting: cooler overcast window light, slightly deeper shadows under the coffee table, same furniture, no extra windows or chairs.Open the workbench and submit your first reference rewrite.
The same workbench supports a range of commercial stills, from packaging lineups to interior restages to sketch-to-finish renders. Each scenario below pairs a visual subject with the creative decision that makes the output usable, so you can see how the tool fits into a real production workflow.
The stills below show real reference-guided outputs and the creative choice behind each one. Reference images and clips on this page are existing site examples, not proof of outputs from the selected model, but they illustrate the kind of decision that produces a usable result. Study the pairing of source and description to understand what the workbench rewards.
Illustrative references. Not verified VideoAny input/output pairs.
Skincare packaging arranged as a reference-guided refresh board shows how a single product shot can spawn a coordinated lineup. An ai photo editor reads the bottle proportions, the cap detail, and the label area, then applies new colorways and surface finishes across the set. The creative decision here is to keep the bottle geometry and the arrangement grid fixed while letting the palette and the label typography evolve. An ai photo editor that holds the structure steady while refreshing the skin gives the brand a consistent campaign language.
Illustrative source for a label-aware packaging rewrite.
A streetwear lookbook frame of an adult in a structured jacket uses the garment shape as the anchor. An ai photo editor can change the fabric, the color, and the background while keeping the pose, the collar line, and the shoulder drop intact. The creative decision is to treat the jacket silhouette as the keep list and the textile and setting as the change list. An ai photo editor that respects the fit and drape of the original garment produces a lookbook variation that still reads as the same cut.
The difference between an acceptable output and a usable one often comes down to how you prepare the source and how you read the result. The three habits below apply to every reference-guided task in the workbench.
01
An ai photo editor reads the entire frame, including background clutter and edge distractions that you stopped noticing. Crop the reference to the exact subject area you want the model to work with, removing anything that should not influence the output. If the job is a product still, crop to the product and its immediate surface. An ai photo editor that receives a tight, clean reference has fewer opportunities to invent objects that were not in the photo.
02
It is tempting to write a long keep list to protect every detail, but an ai photo editor performs better when the keep list is short and the change list is specific. Identify the three or four elements that define the subject, lock those, and let the rest be open to reinterpretation. An ai photo editor that receives a balanced prompt produces a result that feels intentional rather than over-constrained. If the output drifts on a detail you forgot to lock, add it to the keep list in the next run.
03
Completed tasks appear in generation history for preview and download, and the most useful inspection happens when you place the reference and the output side by side. An ai photo editor can reinterpret source details, so check whether the changes serve the brief or distract from it. Look at edges, material transitions, and the parts of the image you did not describe. An ai photo editor that passes a quick glance may still have introduced a texture or a shadow that needs correction.
The answers below cover what to upload, how to judge the result, and what the credit cost looks like. Each question addresses a decision point that comes up during a real session.
The image-to-image workflow accepts a reference image and a description of the desired visual changes. Upload a clear, well-lit still where the subject fills most of the frame, and write a prompt that separates the elements to keep from the elements to change. An ai photo editor reads both inputs together, so a strong reference paired with a vague description will still produce a vague result. Prepare the file and the text before you open the workbench.
The workbench accepts common image formats. Choose a file with enough resolution to show the surface detail and edges you want the model to read. A compressed or low-light source gives the model less information to work with, and the output will reflect that uncertainty.
No. Describe only what should change and what must stay. The reference already provides the composition, so your prompt can focus on the material, color, and finish differences. An ai photo editor uses the reference as the structural base and the prompt as the modification layer.
The workbench has a prompt field, an upload area, and selectors for model and aspect ratio. The three steps below take you from a raw reference to a reviewed result, with checkpoints at each stage so you catch issues before spending credits.
Open the image-to-image workbench inside VideoAny and upload the reference still. Crop the frame to the subject area that matters for the task, removing background elements that should not influence the output. An ai photo editor reads the whole image, so a tight crop is the simplest way to reduce unwanted invention. Confirm the upload looks sharp and well-exposed before you move to the prompt.
In the prompt field, list the elements to preserve and the elements to transform. Be specific about materials, colors, and lighting. Then select Seedream 5 Pro as the model, choose the aspect ratio that matches your output destination, and check the resolution setting. An ai photo editor uses all of these controls together, so a mismatched ratio or the wrong model will undermine even a well-written prompt.
Submit the task and review the result in generation history. Compare the output to the reference and to your keep list. If a material reads wrong or a color shifted, adjust only that clause in the prompt and run again. An ai photo editor rewards targeted iteration; changing the whole prompt resets the learning and wastes credits.
The workbench is ready when you are. Before you submit the first task, freeze the brief, cap the retry budget, and set the fidelity bar you will inspect against. These three decisions turn an open-ended session into a finished deliverable.
An ai photo editor cannot interpret a mood; it reads specific instructions. Write the keep list and the change list in plain, concrete terms before you open the workbench. If the brief says warmer palette, define the target colors. If it says softer shadows, describe the light direction. An ai photo editor that receives a frozen brief produces a result you can evaluate against a clear standard.
Generation uses account credits, and the task cost is shown before submission. Decide how many runs this reference is worth before you start, and stick to that cap. An ai photo editor session can drift into endless refinement if you do not set a limit. When you hit the cap, download the best result and move to the next brief.
Define what close enough means for this task before you see the first output. Is it accurate label placement, consistent material finish, or believable lighting? An ai photo editor produces a range of results, and the fidelity bar keeps you from rejecting a usable output because it does not match an unstated ideal. Write the bar down and check each result against it.
02
A navy athletic sneaker with a pale sole on a plain background is a subject where the silhouette matters more than the colorway. An ai photo editor needs you to separate the keep list from the change list in the prompt. Write that the shoe shape, sole profile, and lacing should remain, while the upper material, color, and background can shift. An ai photo editor that receives a vague description will guess, and the guess rarely matches the brief. Be explicit about the boundaries, and the output will stay within them.
From the uploaded sneaker photo, retain the last, sole shape, stitching, and three-quarter angle. Request a colorway change only: deep burgundy upper, off-white midsole, no extra logos, same catalog crop.Open the workbench and submit your first reference rewrite.
03
A cosmetic glass dropper bottle with flowers on a neutral surface is a finished still that might need a seasonal refresh. An ai photo editor lets you keep the bottle shape and the dropper detail while changing the flower type, the label color, and the surface texture. This is faster than setting up a new shoot and more controlled than generating from text alone. An ai photo editor that accepts a reference preserves the product identity while giving you room to explore variations for a campaign or a catalog update.
Keep the dropper bottle, cap, scale, and front-facing crop from the upload. Do not invent extra label lettering. Move the set onto cool slate with a window wash from the left and a quieter floral accent.Open the workbench and submit your first reference rewrite.
Illustrative lookbook source for a venue or lighting change.
A multi-image sneaker board showing colorway variations on a studio field demonstrates how one reference can drive a full range. An ai photo editor keeps the shoe profile, the sole unit, and the lacing system consistent while cycling through upper materials and accent colors. The creative decision is to lock the product identity and let the surface treatment carry the variation. An ai photo editor that maintains the silhouette across every colorway gives the buyer a clear comparison without distracting shape differences.
Illustrative board for colorway tests from one source shoe.
Creator workflow notes
Working habits for a referenced rewrite, not user reviews
These notes describe common ways to run a keep-and-change pass. They are not user reviews, ratings, or promises about a specific result.
Lighting rewrite
Interior workflow
Keep furniture count and camera height. Change only the window light. If a new chair appears, the change list was too broad.
Colorway rewrite
Product workflow
Keep last, stitching, and angle. Change only the upper color. Inspect the sole and any label area against the upload.
Finish rewrite
Illustration workflow
Keep pose and staging lines. Name the new medium in one clause. Check hands, eyes, and lettering before you accept the plate.
Inspect framing, subject detail, and visual consistency in each result before refining the input. An ai photo editor can reinterpret source details, so a result that looks correct at a glance may have shifted a shadow or softened an edge you needed sharp. Compare the output to the reference and to your written brief, not to an ideal version in your head.
The model can reinterpret source details when the prompt is open-ended or the reference contains ambiguous areas. If the output adds a prop or a texture you did not request, tighten the keep list in your next prompt and crop the reference more closely to the subject. An ai photo editor responds to specificity.
The workbench does not offer region masks. Describe the elements to preserve in the prompt, and the model will weigh those instructions against the reference. If a particular area consistently drifts, adjust the wording to give it more weight in the keep list.
Generation uses account credits, with the task cost shown in the workbench before submission. Submitting a generation task requires an authenticated account. An ai photo editor session is a paid task, so plan your runs around the briefs that need the most control.
The task cost is shown in the workbench before submission and may differ from a text-to-image generation. Check the displayed credit amount before you submit so you can budget your session across multiple rewrites.
Start a separate text-to-image task with a fresh prompt and no reference upload. An ai photo editor is built for guided rewrites; when you need an unconstrained concept, use a different workflow so the reference does not pull the output toward the existing composition.
Joi AI SI
Open the workbench and submit your first reference rewrite.
The workbench is ready when you are. Before you submit the first task, freeze the brief, cap the retry budget, and set the fidelity bar you will inspect against. These three decisions turn an open-ended session into a finished deliverable.
