How to Use Doubao Seedance 4.0 on Imagens.app (Text-to-Image & Image-to-Image)

Doubao Seedance 4.0 is an AI image generator on Imagens.app that supports Text-to-Image (type a prompt) and Image-to-Image (upload a reference image). You can also select an AI model, choose an aspect ratio and resolution, then generate results in seconds. This document explains the workflow, key settings, and practical tips for reliable outputs.

What you can do with Doubao Seedance 4.0

• Text-to-Image: create images from a written description (prompt).
• Image-to-Image (Img2Img): use a reference image to guide composition or style while generating a new result.
• Format control: pick aspect ratio and resolution for the target platform.
• Model-dependent credits: cost is deducted based on the model you choose.

Where to start

1. Open the tool page: Doubao Seedance 4.0 on Imagens.app.
2. Scroll to the creation section (the page includes a Start Creating button that jumps to the generator).
3. Decide your mode: Text-to-Image or Image-to-Image (upload reference).

Step-by-step: Text-to-Image

1. Write your prompt in the prompt box.
2. Select a model (credits are deducted based on the model you choose).
3. Choose aspect ratio (e.g., 1:1 for square posts, 16:9 for headers/thumbnails, 9:16 for mobile stories).
4. Choose resolution (higher resolution usually means more detail and higher cost/compute).
5. Click Generate and wait for results (typical generation time shown on the page is around ~20s, depending on load and settings).

Step-by-step: Image-to-Image (reference upload)

1. Upload a reference image to guide the composition and/or style.
2. Describe what should change (lighting, materials, background, style) and what should stay the same (main subject, framing).
3. Pick model, ratio, and resolution as usual.
4. Click Generate, then iterate with small prompt edits to refine.

Tip from the page: Prompt = subject + setting + style + lighting + camera + details. For image-to-image, the uploaded reference helps keep composition more consistent.

Prompt writing tips (that consistently work)

• Start with the subject: who/what is in the image.
• Add setting: location, background, environment.
• Specify style: realistic, anime, illustration, cinematic, editorial, etc.
• Define lighting: soft daylight, studio lighting, rim light, golden hour.
• Camera & composition: close-up portrait, wide-angle, shallow depth of field.
• Detail constraints: “clean composition”, “high quality”, “sharp details”.

Ready-to-use prompt shortcuts (examples)

The tool page provides prompt shortcuts you can use as templates. Here are common examples (edit them to match your subject):

1) Cinematic Portrait

A cinematic close-up portrait of an original character, soft rim light, shallow depth of field, detailed skin texture, natural color grading, clean composition, high quality

2) Product Shot

A premium product photo of an original wireless earbud case on a minimal background, studio lighting, soft shadows, sharp details, clean reflections, realistic, high quality

3) Image-to-Image Remix

Using the uploaded reference image as the base composition, keep the main subject, refine lighting and materials, increase detail, clean background, premium look, realistic, high quality

4) Fantasy Illustration

A detailed fantasy illustration of an ancient floating island with waterfalls and ruins, golden hour lighting, atmospheric perspective, rich colors, clean composition, high detail

5) Anime Character

An anime-style portrait of an original character, sharp lineart, cel shading, expressive eyes, detailed outfit, soft background bokeh, clean composition, high quality

6) Interior Design

A modern Scandinavian living room interior, natural daylight, soft shadows, realistic materials, clean lines, wide-angle composition, high detail, magazine style

Credits & pricing notes

• Generation is credit-based, and credits are deducted based on the model you select.
• If you want cheaper exploration, start with smaller outputs or fewer iterations, then scale up once the prompt is stable.

Common issues and quick fixes

• Result doesn’t match intent: simplify the prompt, then add details gradually (style → lighting → camera → fine details).
• Inconsistent character details: use image-to-image with a strong reference and specify “same subject, same composition”.
• Artifacts or messy backgrounds: add “clean background, minimal distractions” and regenerate.
• Looks too flat: specify “soft shadows, depth, realistic materials”.

FAQ

Does Doubao Seedance 4.0 support both text-to-image and image-to-image?

Yes. The tool page indicates support for both TextToImage and ImageToImage.

How long does it take?

Typical generation time shown on the page is around ~20 seconds, depending on settings and traffic.

What’s the best way to improve quality?

Use a clearer prompt structure (subject + setting + style + lighting + camera + details), choose an appropriate resolution, and iterate with small changes. For consistency, use a reference image (image-to-image).