Skip to main content
Realtime Gen

Feature Release: Realtime Gen - Generate as you type!

Ayumi Umehara avatar
Written by Ayumi Umehara
Updated over a week ago


Realtime Gen is our second addition to the realtime suite of features on Leonardo.Ai. With Realtime Gen, experience blazing fast text to image generations that evolve as you type. Reiterate, rephrase and refine.

About Realtime Gen

  • Refinement & Personalization: Enhance the overall quality of your realtime outputs with our very own Alchemy Refiner, which not only refines the generated image output but also automatically saves the enhanced output to your Personal Feed for easy access and further customization. Two refinement options are available right in Realtime Generation, Normal and Creative.

  • Style Presets: Switch between different style presets for creative variations of the same prompt.

  • Elements: Style up your images with the familiar Leonardo Elements, now in realtime.

  • Concept Sliders: Change up the age of your character, go dark with the weather, or dial back on the cartoon style and more. All in realtime.

  • Simple UI: Our simplest UI yet, Realtime Gen is made to be straightforward and easy to use.

  • Fast regeneration: Simply click on 'Random Seed' to generate an new image with your current prompt and settings.

  • Advanced Options: Input your own seed values or enable 'Fixed Seed' to keep it consistent.

💡 Leonardo's tip: With the combination of Leonardo.Ai's very own Alchemy Refiner, Realtime Gen works best for photorealistic images. Other styles may work better with the regular AI Image Generation tool.


Realtime Feedback Loop: Generate as you type, and quickly make changes to your prompt and settings based on the live output, all within a few seconds. Rapidly iterate and explore your ideas in realtime.

Enhanced Creative Flow: Focus more on your artistic vision without getting bogged down by additional settings or complex interfaces. Realtime Gen's minimal UI is designed to keep you in the creative zone.

Time Efficiency: Generate more ideas in less time, thanks to the feature's ability to provide realtime feedback on your prompts and various settings, making it easier to achieve the desired result quickly.

Versatility for All Skill Levels: Whether you are a beginner experimenting with basic concepts or a pro drafting up your next prompt, Realtime Gen caters to all levels of creative expertise.

How to use Realtime Gen

  1. Launch: On the main page, click on Realtime Gen in the left sidebar.

  2. Prompt: Input your prompt into the textbox below and watch as the output changes as you type.

  3. Customize: Pick an style preset, throw in some Elements.

  4. Refine: Save your output or use one of the Refine options to create and save both the original image and refined output.

  5. Re-roll: Simply click on Random Seed to get another image.


  • Elements: The weight of each individual element can be adjusted, either in negative or positive values, doing so will affect the image differently.

  • Random Seed: Clicking this generates an new image using the existing prompt and settings. (This changes the seed to an new value even if Fixed Seed is enabled)

  • Fixed Seed: Enabling this sets the seed to an defined value, allowing you to make changes to the settings and prompt without affecting the set seed value.

  • Style Presets: Change up the overall style of your resulting image with our style presets, or pick None if you already have an specific style in mind.

  • Refine Options: Saves your output image to your Personal Feed while also running Alchemy Refiner on it based on 2 different preset options.

    • Creative Refine: Creates an much more smoother output at the expense of style and smaller details such as texture. Works best on features such as faces and hands.

    • Normal Refine: Creates an output more true to the original output, preserving as much of the style and detail as possible.

Concept Sliders

Frequently Asked Questions

How many tokens does it cost to use Realtime Gen?

  • There are no token costs involved for image generation itself. Refining an image costs 8 tokens.

What is the resolution of the output images?

  • Output images are 512×512px (details to be finalized). While the refined outputs are 1496×1496px.


Realtime Gen is in Early Access and as such, we would love to know if you run into any issues or have any other general thoughts about this new feature.

Please feel free to leave your feedback here.

Did this answer your question?