Content drafting
Sign in, bring a content brief from the Content Brief Builder, and get a section-by-section draft: every sentence carries its claim state (bound, gap, stance or no-claim), and where it applies a bound or stance sentence is also marked with the layer its evidence came from; the sentences a human must verify list their references in the verify list; a separate model call checks which must-answer questions the draft covers, and a verify list names what a human must still confirm before publishing.
A GenGrowth sign-in and a content brief are required
The draft takes a Content Brief as its only input — handed over from the brief tool, pasted, or uploaded as JSON — and re-checks its fingerprint before using it. The competitor excerpts and profile facts the brief carries are the only evidence a sentence may cite.
No saved draft, no scheduled run, no content score, no originality claim, and no credit claim. A section whose evidence references fail validation is reported as failed rather than quietly rewritten, and an unfinished coverage check never shows a question as covered.
This tool does not accept a bare keyword. Without a brief there is nothing for the coverage check to measure against.
Bring a brief from the Content Brief Builder, paste its exported JSON, or upload the .json file.
A brief is accepted up to 256 KB; every cross-reference and the fingerprint are checked before it is loaded.
The page accepts only a Content Brief Builder export. The brief is parsed strictly and its fingerprint recomputed; an edited or incomplete brief is refused with the reason, not repaired.
Each writable outline section is a checkbox. Profile facts reach the model only where you allow them: not at all, in the gap-angle section only, or throughout.
Sections are generated in parallel with their own timeouts and a bounded number of attempts each. The model marks every sentence with its claim state and cites competitor excerpts or profile facts where it relies on them; the server validates those references and reports the section as failed rather than rewriting a claim.
With a fresh context, a separate model call reads only the generated sections and the brief's must-answer questions and judges each one covered, partial or not covered. If that call does not complete, coverage is shown as unavailable.
Every sentence carries a claim state; where it applies, a bound or stance sentence is also marked with its evidence's source layer, and the sentences that need human verification list their references in the verify list. Bound sentences are underlined in the colour of their evidence layer, gaps in the error colour; the annotations can be switched off without changing the verify list.
Covered, partial and not covered over the brief's question total. Where they apply, a question shows the section that answers it and the gap the model named; otherwise it shows why its section failed or was skipped. No score.
Sentences with a single competitor source, sentences backed only by your profile, gaps, and stances — grouped, each naming its section and references.
The screen, the copied Markdown and the exported JSON come from the same result. Any section can be rerun on its own; the whole result is replaced and re-fingerprinted.
Once you enter the published URL, the page and the primary keyword are carried to the checker inside this browser, never through the URL bar.
No. Without a brief there is nothing for the coverage check to measure against, so the page accepts only a Content Brief Builder export.
The parser recomputes the brief's fingerprint and checks every cross-reference. A brief edited after export, one whose outline cites a question it does not contain, or one from another schema is refused with the reason shown.
Bound: the sentence cites a competitor excerpt or a profile fact that the brief carries. Gap: the model thought it should be said but found no evidence. Stance: a position from the gap angle, citing profile facts. No claim: a connective sentence that asserts nothing.
The same as the brief's must-answer and outline fields: whitespace-tokenized languages. A brief in Chinese, Japanese, Korean or Thai has no writable section, and the page says so instead of generating.
Neither. It does not publish, write to a CMS or generate images, and the draft lives in your browser tab until you copy or export it. Refresh the page and it is gone.
This page does not show or promise a credit price. Any future price must be defined for this specific tool before the interface claims it.
GenGrowth