Lovision

Model parameters

Configure aspect ratio, quality, output count, and other settings before generating.

What model parameters are

Model parameters are settings you configure before clicking Generate. They control the shape, quality, count, and style of the AI output. Different models expose different parameters — not all options are available for every model.

Common parameters

Aspect ratio

Sets the width-to-height proportion of the generated image or video. Common values:

ValueCommon use
1:1Square (social media, avatars)
16:9Widescreen (presentations, hero images)
9:16Portrait (mobile, stories)
4:3Classic screen proportions
3:2Photography standard
AutoLet the model choose based on prompt

Quality / speed

Some models offer a quality-versus-speed tradeoff:

  • Draft / Fast — quicker generation, lower fidelity. Good for exploring many ideas quickly.
  • Standard — balanced quality and speed.
  • High / Ultra — maximum quality, slower, and typically costs more credits.

Number of outputs

Generate multiple variations in a single run. Setting this to 4 produces four different images from the same prompt. You can review all and keep the best. Each output consumes credits.

Resolution

The pixel dimensions of the output. Some models have fixed resolutions determined by aspect ratio; others let you specify exact dimensions. Larger resolutions cost more credits and take longer.

Background

For image models that support it, controls whether the background is:

  • White or a solid color
  • Transparent (output includes an alpha channel — useful for product shots and icons)
  • Auto (model decides based on the prompt)

Some models can access current information from the web to ground image or text generation in recent facts. Toggle this on when generating content about current events, products, or topics that benefit from up-to-date information.

Style strength / image strength

When using a reference image, this controls how strongly the model follows the reference versus the text prompt. Higher values stay closer to the reference; lower values give the prompt more influence.

Tips

  • Start with Auto aspect ratio and default quality to see what the model produces, then narrow parameters on subsequent runs.
  • Set Number of outputs to 4 on your first run of a new prompt to get a sense of variation before committing to one direction.

On this page