Image Creator

The Image Creator in Oz lets you generate professional, lifelike, and cinematic images fast using either pure text prompts or your own reference images. It’s built for marketers, designers, and creators who need high-quality visuals (product shots, illustrations, logos, sketches, or composites) without a photoshoot or design team.

Here's the flow — simple and powerful:

  1. Choose or Upload Images (optional)

    • Pick one or more images from your Media Library (Oz will surface products and assets it scraped from your site) or upload new files.

    • You can pre-edit any selected image using AI tools (upscale, background removal, recolor, add/remove elements, retouch, etc.). This step is optional — add as many reference images as you like to guide the model.

  2. Write Your Prompt

    • Describe exactly what you want: style, mood, camera angle, lighting, background, and purpose (e.g., social post, product listing, banner). The more specific, the better.

    • You can combine text and reference images so the AI preserves brand details like logos and labels while applying the requested styling.

  3. Select Quantity & Quality

    • Choose how many images to generate and the mix of quality levels (e.g., 1–3 high-quality renders plus several faster standard variants).

  4. Generate

    • Hit Generate and Oz produces your images based on the prompt, selected references, and model settings.

  5. Edit in the Image Editor

    • Generated images open automatically in the Oz Image Editor, where you can:

      • Apply additional AI edits (targeted changes, masks, style transfer)

      • Resize, crop, or add text and overlays

      • Swap elements, replace backgrounds, or refine details

    • When ready, download the final images or post directly to your connected social channels.

Use cases this supports: product photography, realistic lifestyle shots, illustrative art, logos & icons, stickers, editorial sketches, composites, 360-view consistency, and more.

Prompt Examples & Use Cases

  • Artistic Sketch: “Da Vinci-style anatomical sketch of a Monarch butterfly on textured parchment.”

  • Realistic Photography: “Photorealistic close-up portrait captured with an 85mm lens, golden-hour lighting.”

  • Product Shot: “Studio-lit, matte-black ceramic mug on polished concrete, 45° angle.”

  • Sticker/Icon: “Kawaii red panda sticker with transparent background, bold outlines.”

  • Compositing / Targeted Edit: Provide multiple images and ask the model to place a dress from image A onto the model in image B, preserving lighting.

Quick Best Practices

  • Be specific: describe materials, lighting, camera angles, and exact text or logo placement.

  • Give context: tell the model what the image will be used for (ad, hero banner, thumbnail).

  • Iterate: refine the result conversationally — small prompt tweaks produce big improvements.

  • One step at a time: for complex scenes, describe the sequence (foreground, subject, background).

  • Preserve details: explicitly state which elements must remain unchanged (faces, logos, labels).

Last updated