Appearance
Wizard run settings: models, styles, copy, fonts, logo and direction
The last step of the Creative Wizard, Generate and compare, holds a row of six settings tabs. Each tab changes how Nexus builds your batch.
Problem and intended user
The creative strategist (the person who plans and makes ad creatives) wants images that match the brand: the right look, tone, typeface and logo. When those choices are not set, images come back generic or off brand, and each correction becomes another round with an agency or a design tool.
Current workflow
The creative strategist writes these choices into a brief for an agency, or types them into a general AI image tool. The logo and fonts are added afterward in a design tool. To compare two looks, someone makes both and lines them up by hand.
Prerequisites
- A starting point and at least one product. See Generate a creative batch.
- Optional: a logo and fonts in your brand identity. See Brand identity.
How Nexus works
Nexus builds a prompt for each angle and product pair from your angle, scene or job, your products and the six settings. Every model gets that same prompt. AI picks the starting ideas under Surprise me, and AI image models generate the images. The settings change the prompt and the number of images.
The settings lock while a run is in progress. Wait for the batch to finish, change a setting, then press Re-run. Generate and Re-run do not use your credits, but limits apply. You can leave any tab untouched.
Models
Select one or both models: GPT Image 2 and Nano Banana 2. Both get the same prompt, so you can compare their results side by side. You always keep at least one model.
WARNING
Each extra model generates the whole batch again. Two models double the images.
Creative style
Select up to three visual styles, such as Photorealism, Cinematic, Oil painting, Studio packshot, 3D render, Flat vector or Pop art. Select See more for the full list, and Show fewer styles to fold it back. Select a chosen style again to remove it.
With no style, each model picks its own. With several styles, Nexus generates the whole batch once per style, so three styles produce three times as many images as one. At the limit, the panel says "That is the maximum of 3; clear one to swap it."
Copy style
Pick one tone for the words that appear on the image. Select the pill again to clear it. The pills sit in three groups, and each group name carries a short hint:
- Tone (How it sounds.): Funny, Dark humour, Bold claim, Subtle flex, The fact, Conversational, Neutral, Technical, Elevated.
- Voice (Who is speaking.): Testimonial, Founder note.
- Form (How it is laid out.): Rhyme, Question hook, Numbered list, Headline only, Lyrical, Imperative, Long form.
With no tone, each model chooses its own voice.
Font style
Pick the typefaces for any text the model adds to the image. Text that belongs to the product or the logo stays as it is. If your brand identity has fonts, they come first under From your brand identity. The rest are under All fonts. The All fonts heading shows only when your brand identity has fonts.
Use Search fonts to find a family. An empty search shows "No matching fonts." The See … more fonts button shows the full list, and Show fewer fonts folds it back. At the limit, the panel says "That is the maximum of 36; clear one to swap it."
Font messages:
- "Loading your brand fonts…" while the fonts load.
- "Couldn't load your brand fonts." when the load fails.
- "Your brand identity has no fonts set yet" when you set none.
Brand logo
Tick Add my brand logo to these images to attach your logo to every image and ask the model to place it. Clear the box to keep logos off. Left alone, the run follows Embed logo on creatives in your brand identity. The line under the box then reads "Following your brand identity, which currently has this on." or "…off."
Your change applies to this run only. Select Set logo here to add or replace one in Settings > General > Brand Identity. It opens in a new tab. See Brand identity.
Creative direction
Choose where each image starts from.
- Surprise me: Nexus picks a starting idea for every image, so one batch explores several directions.
- Guided: you type one word per angle. If you started from a job, you type one word per product. Nexus still varies everything else, so the images differ.
Type a single word of up to 24 characters in each box. Capital letters are fine. A box with spaces or other symbols shows "Use a single lowercase word — letters and hyphens, no spaces." Generate stays off until you fix it. Blank boxes are ignored.
Images per model
Use the + and − buttons beside Generate to set images per model. The total follows this rule:
images per model × models × angle and product pairs × styles
Nexus keeps a batch between 4 and 60 images. The largest value you can pick shrinks as you add products, models or styles.
Generate and re-run
Press Generate to start the run. After the first run, the button reads Re-run. Change any setting, then press Re-run to start a new generation. Generate and Re-run do not use your credits, but limits apply.
Exact output
The output is a row of six tabs and a batch built from them. Each tab shows its current value under its name. Creative style is open when you arrive. Only Models is required, and it carries an asterisk.
The six tabs
| Tab | What it sets | Default |
|---|---|---|
| Models | Which image models generate the batch | GPT Image 2 |
| Creative style | How the images are rendered | Any style |
| Copy style | How the words on the image are written | No tone |
| Font style | Which typefaces the words use | Any font |
| Brand logo | Whether your logo goes on the images | Your brand identity setting |
| Creative direction | Where each image starts from | Surprise me |
The Generate button
The Generate button names the image count, for example Generate 8 images.
What to do with it
| What you see | Decision or action | Owner |
|---|---|---|
| Two model columns for one prompt | Compare them, then keep the images you prefer. | creative strategist |
| The image count on Generate | Choose the models and styles you want, then press Generate or Re-run. | creative strategist |
| Set logo here | Add or replace the logo. It opens Settings > General > Brand Identity in a new tab. | ecommerce manager (the person who runs the store) |
| Images from a finished run | Review them and choose Use selected. See Review and use batch results. | creative strategist |
Expected value and measurement
The settings can help you get images closer to your brand and compare looks and models on one prompt, without a new brief for each trial.
To check, compare the model columns and the batches in Create > Creative Batches. After the ads run, read Spend and ROAS for them in Performance > Ad performance. Change one tab per Re-run, so you can tell which setting made the difference. When you change several settings, or run other offers or budgets at the same time, you cannot isolate the effect.
Limitations and exceptions
- The picker offers two models. You can pick up to three styles, one tone and up to 36 fonts.
- AI image models write the text and place the logo. They can get the words or the logo placement wrong, so check each image.
- The Brand logo box is unavailable in three cases. It says "Loading your brand identity…" while the data loads, "Couldn't load your brand identity." when the load fails, and "Your brand identity has no logo yet." when your store has no logo. It also stays unavailable while a run is in progress. Set logo here opens Settings > General > Brand Identity.
- A Guided word that is not a single lowercase word shows "Use a single lowercase word — letters and hyphens, no spaces." Generate stays off until you fix it.
- If the total is out of range, a message beside Generate says why. Choose fewer products, models or styles.
- In Demo Mode a banner reads "The Creative Wizard does not generate in Demo Mode". You can set every option, but the batch does not start.