Measure
Two weeks after every change, clicks and position are compared against that page's own pre-change baseline, and the result is recorded as a win, a loss, or flat.
Features
Fix what you have. Protect what works.
FOXQUILL refreshes the posts you already rank for and writes new ones to current SEO and GEO standards, in each blog's locked voice, on autopilot at the level you set. Here is what runs under the hood.
QA gate - draft #27
publication-readyGrounding0 uncited stats
Structureanswer-first, FAQ, schema
Voice8.5 vs locked anchors
Linksinternal both ways
Weak draft?
It regenerates until it passes
Diagnosis
How does FOXQUILL decide what to fix?
It scores every post from your own Search Console and Bing numbers, not a third-party scrape, and sorts each underperformer into one clear action. Five opportunity types drive the queue:
- Striking distance: position 8 to 20 with real impressions, where a small lift moves a page onto page one.
- Low CTR for position: high impressions but clicks below the curve, usually a title or intent problem.
- Decay: clicks sliding over recent months against the prior period.
- Query gaps: searches the page gets impressions for but never actually covers.
- Cannibalization: two URLs splitting one query, flagged to consolidate, not double-refresh.
Brand voice
How does each blog keep sounding like itself?
Voice is generated once from a persona, shown to you as real sample paragraphs, then approved, locked, and versioned. Every rewrite reads that locked profile and its anchors as the style reference. When an article lands short, the editor's critique becomes a lesson this blog's writer applies from the first sentence next time, so the voice gets sharper with every piece and a whole network of blogs never collapses into one generic tone.
AI search
How is the content built to get cited by AI engines?
Rewrites lead with the answer, use question-shaped headings, add a real FAQ, and ship clean JSON-LD, so a model can lift a clean answer without guessing. The same craft wins classic rankings and AI citations at once, and every publish pings Bing the same hour, because Bing's index feeds ChatGPT and Copilot.
Quality gate
What stops a weak or invented draft from shipping?
A two-layer gate. Deterministic linters check structure, metadata, grounding, and coverage for free; then AI judges score voice, answer-first quality, and fact attribution. Research keeps only facts whose values appear in the source, and a judge verifies each number sits on the right subject, so a statistic can never drift onto the wrong thing. A draft that misses regenerates until it is clean or gets flagged for you, never auto-published.
Repair, not restart
What happens to a draft that just misses the bar?
It gets improved, not thrown away. One focused pass fixes only what the gate flagged, reusing the facts already researched, in the same locked voice. If the pass cannot beat the current draft it leaves it untouched and tells you what is left, and a draft that clears the bar earns its images then. You skip the old reject-and-rewrite-from-scratch over a single rounded number or banned word.
Internal linking
Does it find internal links and duplicate pages?
Yes. Every post is embedded into a shared vector index, so a rewrite adds relevant links out, updates older posts to link back, and catches two pages competing for the same query before they hurt each other. That builds topical clusters search engines and AI models both reward.
After publish
Publishing is the halfway point
Most tools stop at the draft. FOXQUILL keeps watching every page it touched and closes the loop, so autonomy is earned with measured wins, never assumed.
Learn
The editor's critique becomes a per-blog lesson the writer applies next time, and the QA judges keep the voice anchored as the profile sharpens.
Act
Pages that lost ground go back into the fix queue on their own, and a page that never got indexed is resubmitted to Bing, so nothing weak sits unnoticed.
Everything inside
The full spec
Reads your network
- WordPress REST sync of every post, embedded into a shared index
- 16 months of Search Console history per URL, refreshed twice a day
- Bing Webmaster performance as its own scored signal
- AI-citation tracking (who quotes you) over time
Writes to standard
- Answer-first structure, question H2s, FAQ, comparison tables where they fit
- Length-checked SEO title and meta description as first-class outputs
- Best-fit existing category, lean controlled tags, no tag sprawl
- On-brand images generated per article, added post-generation
Protects what works
- Every write is a reversible WordPress revision, one-click rollback
- Retire safely: 301 or consolidate to keep link value, delete only true junk
- No fabricated stats, quotes, or prices, enforced by code
- Staggered rollout, proven on one blog before the network
Runs at your pace
- Three autonomy levels per blog: suggest, auto-draft, guarded publish
- Guarded publishing unlocks only after measured uplift on that blog
- Schedule approved drafts across the week, move or cancel any of them
- Multi-tenant: each user sees only their own blogs and data
See it on your own blog
Tell us about your network and we will map the first blog's opportunities with you, from your real Search Console data.
Prefer the walkthrough first? See how it works.