See Mode

See Mode

See mode in Questsmith lets you create and add images to your adventure during gameplay.

We do not claim copyright on any user-generated images. You own the pictures you generate, so download your favorites and use them how you want.

Creating images usually requires Credits, unless your subscription tier or image model says otherwise.

Getting Started

Open the settings sidebar through the gear icon in the upper right corner. Go to the Gameplay section and the AI Models tab, then scroll to the Image Generator accordion and set the image source to the model you want.

Back on the game screen, choose Take a Turn, open the input mode options with >, and select See. Type what you want to see, then use the picture icon on the right side of the input bar to submit the prompt.

If you leave the prompt blank, See will automatically build a prompt from your recent actions, story cards, and plot essentials.

Image Options

After an image is generated, click it to open the Image Options popup.

• Share: Create a shareable link that can be opened in a new tab and downloaded.
• Download: Save the image to your device.
• Delete: Remove the image from your adventure. Deleted images cannot be restored, and the exact same image will not be generated again.
• Retry Image: Generate a new version using the same credit cost as the original image.
• Edit Prompt: Add more detail before retrying so the AI has better context.

Undo and Redo from the top of the game screen work for images just like they do for text inputs.

Prompt Basics

Practice with simple one- or two-word prompts first. Prompt text can include Unicode, emojis, slang, misspellings, and foreign languages, although English usually works best.

Capitalization is ignored, but punctuation can help separate modifiers, such as blue ball, red brick versus blue ball red brick.

Create Complex Images

Once you are comfortable with simple prompts, you can move into more advanced prompt engineering.

Basic Prompt Structure

Start with the kind of image you want, then add modifiers, foreground subject, background, style, quality, and artist inspiration.

A useful format is: [Medium] of [Subject] in [Style] with [Details] by [Artist(s)].

• Medium: watercolor painting, photo, portrait, illustration.
• Camera angle: close up, overhead view, view from the river.
• Adjectival modifiers: dramatic, red, stone.
• Foreground: ship, castle, dragon.
• Background: in the forest, on a beach, in a city.
• Style: art nouveau, minimalist, dark and foreboding, Pixar style.
• Quality: beautiful, sharp, intricate, award winning.
• Artist: by van Gogh, by Donato, by Tyler Edlin.

List the most important information first. The beginning of the prompt matters more than the end if the prompt gets long.

For photo-realistic images, always mention the lighting, such as dramatic, high-key, or low-key lighting.

Influence the Final Output

Repeating important words in different forms can help the model focus on the subject. For example, mechanical, robot, android makes it more likely the image will emphasize an android.

Adding descriptive elements like highly detailed, surrealism, sharp focus, matte, elegant, ambient lighting, epic composition, or 4k can also push the result in a stronger direction.

The model is better when you describe a situation instead of just an expression. It also struggles with hands and dinosaurs, so prompts that avoid hands often work better, while dragons tend to look stronger.

Useful Tips

• Use the beginning of the prompt for the most important subject and medium.
• Keep prompts under 400 characters when possible.
• Mix multiple artists or styles when you want a more original result.
• Use the image generator as another way to customize and control your gameplay experience.