A model will draft anything you ask, fluently. It will also make you sound like every other page a model wrote: the placeless voice nobody chose. DOMAINS is the fixed point that keeps that from happening. It reads where your draft sits across seven dimensions of voice, and it shows the one change that moves it toward the voice you want, then holds the draft to that target the same way every run. It never rewrites for you. You keep the pen.
The reflex right now is detection: platforms scan your writing to guess whether a machine was involved, and answer with a probability that flags human writers as often as the machine. That is the small question, and it is an arms race no one wins. The real question is not who typed the words. It is whether they sound like anyone at all. doloop measures that, the same read whether you wrote the draft or your model did. And because it hands you the one change rather than the finished text, you stay in the draft, editing and reshaping, instead of skipping to the end. Writing is still thinking. This keeps you in it.
DOMAINS is doloop’s writing tool. It reads your draft across seven dimensions of style, paragraph by paragraph. Diction, how specific your words are. Orchestration, how much your sentence lengths vary. Mechanism, how you build cause and effect. Addressivity, how often you speak to the reader. Imagery, the physical detail you carry. Narrative, how much the text moves. Suppression, how spare you run. Those seven initials spell the name. Each one is simple. Together they make the shape that is yours.
How much your writing varies as it goes: concrete against abstract, short sentences against long, scene against reflection. A draft that stays in one gear reads flat. The profile across all seven dimensions is where your voice lives.
DOMAINS names the human register your prose sits closest to, from personal narrative to legal opinion, and measures how far each other register is. You see, in one shape, the company your writing keeps.
Pick a target register. DOMAINS shows the single dimension pulling you away from it and lights up the exact words where that dimension lives in your draft. You act on one thing, not a wall of notes.
DOMAINS is deterministic. The same draft scores the same way every time, so a second look never moves the number on its own. You bring your own model; the reading does not shift under you.
Your seven scores draw one shape. The register you aim for draws another over it. The gap between them is the work, and it is right there to see.
Every word that carries a dimension lights up in its colour. Hover one and it tells you what it is doing, and whether it pulls you toward the voice you picked or away from it.
Score your draft → The same content, every voice → See the evidence
Call DOMAINS on its own, or run it inside the doloop machine (our hosted service), which keeps your prior reads and the target you set. Same reading either way. The difference is whether the run carries state.
Send a block of text, get its seven scores, its register, and the distances back. Nothing to set up, nothing kept between calls. The fast way to read one draft.
Connect your own model to the doloop machine in writing mode. Your model drafts, DOMAINS reads each one against the voice you set, and the shape converges toward that voice over successive passes.
Want this on your writing pipeline? Talk to us, or see the other surfaces.
Routes you to a real doloop page, asks when your question is ambiguous, or tells you when there is no answer. No model runs on the answer path, so it cannot invent one.