Image workflows are runtime-gated like the rest of the platform: provider keys, build profile, and registration mode decide what the client can actually call.
SindByte can generate and refine web assets from the same local automation surface that handles files, text, and deployment steps. The image path is especially useful when you want an agent to create a full page package instead of stopping at raw copy.
CUTools::GenerateImage
CUTools::PictureResize
PictureConvert
PictureEdit
OpenAIImageToolEnabled=true
OpenAIImageModel
gpt-image-1
1) GenerateImage(prompt="clean MCP capability map", target="file", target_file="assets/img/feature-map-generated.png") 2) PictureResize(source="file", source_file="assets/img/feature-map-generated.png", target="file", target_file="assets/img/feature-map-generated-web.png", width=840, height=840) 3) PictureConvert(source="file", source_file="assets/img/feature-map-generated-web.png", target="file", target_file="assets/img/feature-map-generated-web.jpg", format="jpg") 4) Integrate the generated asset into HTML/CSS and continue with normal site edits
Write the asset directly into the site or build workspace when the next step is HTML/CSS integration.
Keep generated output inside the workflow state when you want an agent to compare variants before writing anything to disk.
Move assets quickly into another desktop tool or design step without saving intermediate files.
Use when you want a durable file plus a transient workflow copy at the same time.
Combine image generation with file writes, HTML edits, and deployment steps through the rest of the SindByte stack.