Recraft v3 Design Image Generation

Problem It Solves

When producing brand visuals, illustrations, posters, packaging, or UI assets, teams often need more than “generate one image.” They need a controllable design style, reference-image fidelity, and a result that can flow into a pipeline. recraft v3 targets this scenario by wrapping Recraft v3's design-image capability as a command for agents or scripts, reducing manual parameter assembly, asset uploads, and task-status tracking.

How It Works

• The core command is dlazy recraft-v3, which sends prompts and parameters to the dLazy API for inference.
• It supports text-to-image for prompt-driven design images and image-to-image for local reference images uploaded to dLazy media storage.
• Authentication uses a dLazy API key, stored locally or supplied via DLAZY_API_KEY; keys can be rotated or revoked per organization.
• Async runs can return a task payload first, then be polled with dlazy status --wait; errors surface as auth, missing parameter, local file, balance, or backend failures.

Boundaries

This skill is a cloud SaaS client, so it is not suitable for offline generation, strict private storage, or local model inference. Local files are uploaded through dLazy, and output URLs are hosted in the cloud. Before use, verify API key scope, account balance, service terms, and brand compliance for generated assets.