Volcano Ark Doubao Seedream Text to Image

Problem

When calling Volcano Ark Seedream models, developers often deal with the same small pain points: whether the API Key belongs to Ark, different endpoints for text-to-image and image-to-image, and easy-to-miss parameters such as model ID, prompt, and size. This skill wraps those API calls into repeatable CLI and interactive flows, which is useful for prompt testing, generating images locally, or adding an image generation step to a local workflow.

How It Works

• Text-to-image: generates images from a text description, using parameters such as --prompt, --model, --size, and --output.
• Image-to-image: sends --input-image plus an editing instruction to modify an existing image.
• Model selection: supports versions including Doubao-Seedream-5.0, 4.5, 4.0, and 3.0; the default model is doubao-seedream-4-0-250828.
• Interactive mode: guides you through choosing a model, selecting text-to-image or image-to-image, entering a prompt, and generating the image.
• API calls: text-to-image uses POST /images/generations, while image-to-image uses POST /chat/completions, under the Volcano Ark base URL.

Before first use, the API Key should be configured and should have access to the intended image generation model.

Boundaries

The skill is focused on image generation and image editing API calls. The provided materials do not show support for video, 3D, or other generation tasks, and it does not handle prompt-effect tuning directly. Supported sizes include 512x512, 768x768, 1024x1024, 1024x2048, and 2048x1024; generation time is usually around 5-30 seconds depending on server load. If you see an invalid API Key or unknown model error, first check that the Key is from Volcano Ark and has permission for the target model.