Edit Workspace

Opening the editor

Click any generated image to open it in the Edit Workspace. From here you can apply AI-powered edits without losing the original asset. Every edit creates a new version, so you can always revert to any previous state.

The Edit Workspace is available for all image assets — whether generated on Gizmoji or uploaded from your device. Non-image assets (video, audio, 3D) are downloaded and managed from their detail view.

Upscaling

Increase the resolution of any image using AI-powered upscalers. This is essential when you need high-resolution output for print, large-format display, or downstream production use.

• Budget upscalers — Fast and affordable. Good for quick resolution bumps where fine detail is less critical.
• Value upscalers — Balanced quality. Adds detail and sharpness while maintaining the original character.
• Premium upscalers — Highest quality. Reconstructs fine details, textures, and edges for professional-grade output.

Most upscalers support 2x and 4x scaling. The credit cost is displayed before you confirm, so you can choose the right balance of quality and cost.

Restyling

Transform the visual style of an image while preserving its composition and subject matter. Describe the target style in a prompt and the AI re-renders the image accordingly.

Example restyle prompts:

• “Studio Ghibli anime style, soft pastel colors”
• “Photorealistic, cinematic lighting, shallow depth of field”
• “Oil painting, impressionist brushstrokes, warm tones”
• “Cyberpunk neon aesthetic, rain-soaked streets”

The strength parameter controls how much the output deviates from the original. Lower values keep more of the original composition; higher values give the AI more creative freedom with the new style.

Image-to-image iteration

Use image-to-image generation to refine specific aspects of an output. This is different from restyling — instead of changing style, you're directing the AI to produce a variation with specific changes.

• Adjust the prompt — Add or change details in your text prompt to guide the variation.
• Control strength — Low strength keeps most of the original; high strength allows major changes.
• Try different models — Each model interprets prompts differently. Switching models on the same reference can produce surprisingly different results.

This iterative workflow is how professionals converge on a final result: generate a base, iterate 2–3 times to dial in details, then upscale the winner.

Comparing versions

As you generate variations and edits, the Edit Workspace keeps every version accessible. You can:

• Browse all versions in the version history panel.
• Compare two versions side by side to evaluate differences.
• Promote any version to the approved/active version.
• Download any individual version without affecting others.

Nothing is ever overwritten — every generation and edit is preserved as a distinct version with its own ID.

Approving frames for video

Before you can use an image as input for video generation, it must be approved. This is a deliberate quality gate that prevents you from spending video credits (which are typically higher) on a frame you haven't reviewed.

To approve an asset:

1. Open the image in the Edit Workspace or asset detail view.
2. Review it and make any final edits or upscales.
3. Click the Approve button to change its status from Draft to Approved.
4. The image is now available as input in the Video Studio's image-to-video mode.

Recommended editing workflow

For the best results, follow this workflow:

1. Generate — Create a base image with a budget model to test your prompt.
2. Iterate — Refine using image-to-image with adjusted prompts and strength settings.
3. Upgrade — Once the composition is right, regenerate with a premium model for higher quality.
4. Upscale — Increase resolution for final output or print.
5. Approve — Mark the final version as approved for use in video or delivery.

Frequently asked questions