Creating an Image with BLNG

There are different ways to generate an image in the app, depending on how you combine three variables:

Canvas input – you can draw something, sketch, or upload an image
Text input – describes what you want to see
Canvas influence – controls how much the canvas affects the result

Each generation starts with a mix of these. You can go simple (just a prompt), or combine everything for more control.

Basic Approaches

Text to Image Type a prompt and generate an image from scratch. Canvas must be empty or canvas influence should be set to 0%.
Doodle to Image Draw something quick and loose. Use low canvas influence to let the model build creatively around it. Add a prompt for context.
Sketch to Image Make a structured or detailed base. Use high canvas influence to retain layout and form. Add a prompt to shape content and style.
Image to Image Upload an image as your base. Use canvas influence to decide how much of it is preserved. Add a prompt to guide changes.

Pro-Tip

There’s no single way to create — the real power comes from how you combine the tools. Start simple, test variations, and adjust as you go. Whether you’re sketching from scratch or remixing existing ideas, each generation is a new opportunity to shape your vision.

Combining Variables: Canvas, Text & Canvas Influence

Once you’re familiar with the basics, the next step is to combine these three variables to get more precise or interesting results:

Canvas input – draw something or upload an image
Text input – describe what you want to create
Canvas influence – adjust how much the canvas affects the output

This is where you can start experimenting — sometimes giving the model more freedom (with a lower canvas influence) will surprise you in a good way. Other times, if your drawing or image is already clear and intentional, a higher influence might work better to retain your composition.