Jimeng Text-to-Image

What It Solves

When you need text-to-image generation from a prompt, but do not want to assemble HTTP calls, handle authentication, upload assets, or poll async jobs manually, jimeng t2i gives an engineer-oriented CLI entry point. It reduces image generation to a repeatable flow: provide a prompt and parameters, then receive an output URL or an async task state. That makes it easier to embed in scripts or agent pipelines.

How It Works

The skill is a thin client over the hosted dLazy API, with Jimeng as the image model. The usual flow is:
- Configure a dLazy API Key; it can be stored locally or supplied per invocation.
- Run dlazy jimeng-t2i and send the prompt and generation parameters to api.dlazy.com.
- If local file paths are supplied, the CLI uploads them to files.dlazy.com so the model can read them.
- Synchronous calls return image URLs; with --no-wait, the response includes generateId and status, which can be polled until completion.

It also supports pipe-style references, allowing values from upstream outputs to be passed into later steps.

Boundaries and Notes

This skill depends on a cloud SaaS service and does not run local inference. A valid dLazy API Key is required, and you should account for balance, organization permissions, and service terms. Errors may be caused by unsafe prompts, missing parameters, unreadable files, service failures, or async execution failures. Generated assets and uploads flow through dLazy endpoints and media storage, so sensitive content should follow the platform's compliance requirements.