Text that contradicts an image is a no-op.

A modern generator takes several inputs at once — a subject reference, a style reference, a first frame, a pose map, the text — and they are not peers. Each one governs some visual attributes, merely influences others, and ignores the rest. Writing “in a red jacket” while supplying a reference wearing a grey one does not blend, does not average and does not win. It does nothing at all, and the sentence sits in the prompt looking like it worked.

The engine is free and needs no account. Paste what you are asking for and what you are supplying, and every attribute is resolved to the channel that decides it. Two governors is reported as a tie rather than guessed at — which one a pipeline keeps is a property of that pipeline, not of your request. It also names the attributes nothing governs, which are re-rolled every generation and read as model instability, and the references that govern nothing you asked for. Adding a first frame to a request that already has a subject reference feels like it can only help; it ties on four attributes and silences two images. Every lane ships a worked example that costs nothing. Writing the report needs credits.