Skip to main content
Skip to content
Generate an image from text

Generate an image from text

Choose a model and visual style in Playground, write a prompt, and generate an image.

Before you start

Open Generate images. Image generation is on every plan. Seedream 5.5 is marked Best Quality. The Create button shows the credit total. Its accessible name is Create image.

Steps

  1. Choose Image. If the panel is collapsed, expand it. In a project, the prompt says Describe your image, use @ for elements. Outside a project, it says Describe your image.

    Image generation panel with a sample fox prompt, Seedream 5.5 at 1.5K, and a Create button showing 4 credits.
  2. Open Image aspect ratio and choose the frame. Open Image model and choose a model. Seedream 5.5 is marked Best Quality. A model that accepts pictures shows Supports image references. Then set Resolution or Quality when that model offers it. Seedream 5.5 uses Resolution: 1K and 1.5K are 4 credits, and 2K is 5.

  3. Open Director for framing, camera, lighting, or style presets. The image panel has no separate Visual style button. Set the project’s visual style in Project settings.

  4. Write the subject, action, setting, and framing. For example: “A curious red fox exploring an empty film set at night, wide shot, pools of warm light.” Add reference images only when the model shows Supports image references.

  5. Read the credit total on Create, then choose Create. Wait for the result. Choosing Create again starts another paid image.

  6. Find the completed image in the Playground gallery. Open it to inspect the result, or use its download control to save a local copy.

Expected result

The image appears in the Generate images gallery and in My Assets → Images. Choose it from My Assets when you add it to a project.

Tips

Start with the result you want rather than a long list of conflicting styles. Model, resolution, and reference support can change the available options and cost; use the values shown in the app.

Documentation