This Skill encapsulates the Volcengine Seedream-4.5 image generation capability, supporting the generation of high-quality AI images through text prompts or reference images.
Usage
Text-to-Image
Generate a single image (default resolution):
uv run {baseDir}/scripts/generate_image.py --prompt "一只赛博朋克风格的猫" --api-key YOUR_API_KEY
Specify resolution (e.g., 2K, 4K, or specific pixels):
uv run {baseDir}/scripts/generate_image.py --prompt "壮丽的山川日出" --size "2K" --api-key YOUR_API_KEY
Image-to-Image
Provide a reference image URL:
uv run {baseDir}/scripts/generate_image.py --prompt "将其风格变为印象派" --image "https://example.com/input.jpg" --api-key YOUR_API_KEY
Generate Image Group
Generate a group of content-related images (up to 15 images):
uv run {baseDir}/scripts/generate_image.py --prompt "一个宇航员在不同行星上的探险经历" --sequential --max-images 4 --api-key YOUR_API_KEY
API Key Configuration
You can provide the API Key in the following ways:
- Command-line parameter
--api-key - Environment variable
VOLC_API_KEY
Parameter Description
--prompt: (Required) The text description for image generation.--model: (Optional) Model ID, defaults todoubao-seedream-4-5-251128。--size: (Optional) Image size. Supports2K,4Kor2048x2048and other formats.--image: (Optional) Reference image URL or Base64 encoding.--sequential: (Optional) Enable the image group generation feature.--max-images: (Optional) The maximum number of images to generate in an image group (1-15).
Workflow
- Call the
generate_image.pyscript. - The script will output image links starting with
MEDIA_URL:. - Extract the links and display them using Markdown syntax:
。 - Unless requested by the user, there is no need to download the images.
Notes
- Seedream-4.5 supports Chinese and English prompts.
- The image group feature is only effective in the Seedream-4.5/4.0 models.
- Ensure that the provided image URL is publicly accessible.