Every plan starts with 14 days of full access. No card up front, no per-word metering.
For one writer with one voice to protect.
For a small team sharing a byline.
For agencies managing many voices.
For marketing departments at scale.
In build, not yet included: Website & social ingestion · Brainstorming from your calendar · API access
No, and that is deliberate. We tested it: a detector-driven verdict marked a damaged article green and a client-approved one red. You get a report instead, showing where a passage sits against human ranges, against your own approved work, and which hard rules it breaks. One number would hide exactly the thing you need to see.
One client, built from writing they have already approved. We measure their range, then every later draft is compared against it rather than against a generic idea of good writing. A client with two audiences gets a profile for each.
They are third-party claims, shown in each vendor's own terms and dated, never merged into a verdict. Detectors disagree with each other and retrain without notice; we have watched the same text go from 41 to 99 overnight. We record them and never promise a pass.
It is used only to build your own voice profiles. We do not train models on your writing and we do not pool it across clients.
Yes. Every draft exports to Word, and the editorial calendar is yours.
Agencies reselling DeftCopy to their own clients get volume rates. Talk to sales.