SEO Agent
Optimization is
a gate, not a grade.
Most tools give your content a score and publish anyway. The SEO Agent's checks are pass/fail: claims that can't be sourced get rewritten or removed, and articles that read like a paraphrase of page one don't ship.
Gate one
The fact-check: every claim, a source.
Every factual claim in a draft is extracted, checked against a source, and mapped back to its exact position in the text. You see the report in the review screen — highlighted claims, verdict colours, the source for each. Unsupported claims trigger one targeted rewrite; still unsupported means the article fails.
Treatment report
Simulated democlaims extracted: 12
verified: 11 · rewritten: 1
each claim → verdict + named source
Gate two
The distinctiveness gate: say something page one doesn't.
Each article is scored against what already ranks and against your own published posts. Below threshold, it gets one rewrite attempt; if it still can't justify its own existence, it is killed — visibly, with the run marked failed, never silently published as filler.
On top of the gates: title and meta description written to length, and heading structure derived from the actual content. roadmapFAQ blocks and schema markup generation are on the roadmap and not claimed as live.
Distinctiveness
vs live SERP context: scored 0–100
vs your own corpus: no self-cannibalising
below 60 → rewrite once, then fail
Start with the free check-up.
Put your website address in, and in about a minute you get a report: how fast your site loads, whether Google can read it, and which customer searches you have no page for. No card, no call.
Takes about a minute · No card, no call · You're always in control of what publishes