Your prompt and settings follow you through sign-in. Review them before generating.
Start with a visual direction
Explore creative references, reuse a scene description and add your own ideas.
White tiger
A majestic white tiger among snow and mist, detailed fur and a cold mountain atmosphere.
Noir portrait
A noir portrait under a dark hat, one narrow beam of light shaping the face against black.
Alpine gateway
A grand stone gateway overlooking a snowy alpine valley, reflected light and deep perspective.
Giant tree
A lone person beneath a colossal dark green tree, soft light at the horizon, epic scale.
Creating with Imagen 4
Imagen 4 in LoveImage is a text-to-image workflow with a Standard output setting. Approach it like briefing a photograph: identify the subject, location, light and viewpoint. It is a practical option to explore photographic scenes without starting from an uploaded picture.
Describe the physical scene
Specify where the light comes from and what the main surfaces are made of. Concrete material and lighting descriptions are easier to evaluate than a string of quality adjectives.
Use framing to direct attention
Choose the aspect ratio for the intended placement, then explain shot distance and the subject’s position. Inspect edges and background distractions as carefully as the main subject.
How to get more from Imagen 4
Write a photographic brief with concrete scene details, then judge the result like a photograph.
Set up a photographic scene
For a still life, landscape or lifestyle concept, decide the subject, environment and time of day. Make the intended use clear so the composition leaves the right amount of space around the subject.
Describe light and materials
Explain where the light comes from, whether it is soft or hard, and how the main surfaces should look. Add camera distance and viewpoint before relying on broad adjectives such as cinematic or premium.
Look closely at physical details
Check shadows, reflections, contact points and repeated textures. Inspect any hands or written text separately. Revise the specific physical detail that feels wrong rather than adding more quality adjectives.
Imagen 4 use cases and prompt ideas
Start with a concrete task. These original prompts are starting points to adapt. Add one to the composer above, then review your settings before generating.
Product still life
Explore a photographic presentation for an object. Check contact shadows, reflections and shape before treating the result as a usable marketing asset.
A clear glass perfume bottle without a label on a pale stone surface, soft morning window light from the left, a subtle reflection, quiet cream background, close-up product photograph.
Travel editorial concept
Set a location, weather and viewpoint for a coherent scene. Treat generated locations as visual concepts unless you have independently checked their factual accuracy.
A quiet coastal café terrace at sunrise, empty wooden chairs, white linen tablecloths, calm blue sea in the distance, eye-level travel photograph, natural warm light, no people or text.
Imagen 4 at a glance
What this model actually accepts. The studio shows exactly these options when it is selected — nothing more, nothing it would reject.
Explain where the light comes from, whether it is soft or hard, and how the main surfaces should look. Add camera distance and viewpoint before relying on broad adjectives such as cinematic or premium.
STEP 02
Choose available settings
Available here: Default output (no resolution selector). These are the settings enabled in this workspace, not an exhaustive list of the provider’s options.
STEP 03
Review the result
Check shadows, reflections, contact points and repeated textures. Inspect any hands or written text separately. Revise the specific physical detail that feels wrong rather than adding more quality adjectives.
Imagen 4 or Seedream 5 Pro?
Compare Seedream 5 Pro for another interpretation of the same scene and access to its available output tiers. If your task requires changing an existing photograph, use a reference-capable model instead of either text-only selection.
Model settings currently available in LoveImage
Feature
Imagen 4
Seedream 5 Pro
Input
Text
Text
Output tiers
Default output
1K / 2K
Aspect ratios
1:1 / 4:3 / 3:4 / 16:9 / 9:16
1:1 / 4:3 / 3:4 / 3:2 / 2:3 / 16:9 / 9:16
These are the settings exposed in this integration, not a performance ranking. Resolution can affect duration choices, and matching tier names do not imply equal quality. Review the editor’s configuration and credit estimate before submitting.
Imagen 4 in LoveImage is a text-to-image workflow with a Standard output setting. Approach it like briefing a photograph: identify the subject, location, light and viewpoint. It is a practical option to explore photographic scenes without starting from an uploaded picture.
Available here: Standard. These are the tiers exposed in the LoveImage editor. Pixel dimensions depend on the model and aspect ratio; a quality-tier name is not necessarily a fixed pixel size. Check the downloaded file before delivery.
Current options: 1:1 / 4:3 / 3:4 / 16:9 / 9:16. Match the frame to the intended placement and describe the subject position and empty space in your prompt. Where auto is offered, it uses automatic framing. Check the crop before refining details.
This selection uses text input and does not accept a reference image. For a photo-led result, choose a reference-capable model in Image to Image or Image to Video.
Describe the subject, spatial relationships, framing, lighting and materials in that order. Put requirements first and quote any short text exactly. Change one main variable between drafts before adding more detail.
Inspect subject shape, hands, edges, reflections, text and numbers at full size. Then check readability and spacing at the intended display size. Compare branded or product imagery with the original source before using it.
Credits depend on the model, size, duration and output count. Review the total shown in the editor before submitting; a starting price in the specifications does not cover every configuration. Check the estimate again after changing settings.
The no-sign-up image trial uses Z-Image; opening this model page does not change the trial model. To use the model selected here, sign in and meet its plan and credit requirements. Review the editor before submitting a generation.
This selection accepts a text prompt, not an uploaded reference. Open Image to Image and choose a model that supports reference input when the original photo needs to guide the result.
Review reflections, shadows, object contact and repeated textures individually. Rewrite the problematic detail in concrete terms; photographic styling does not ensure that every physical relationship is correct.