Prompt-targeted
Each article exists to win one tracked prompt.
Publish the page that answers it.
Takes a buyer question you are losing and drafts the long-form page that wins it, with sources, an outline and a rubric pass. You review, approve and publish.
Free analysis · 5 credits · no card · first score in about 30 seconds
one prompt per draft · evidence attached · you publish
target "coinbase alternative for spanish users"
length 2,400 words · 6 sections · FAQ block
includes fee table · MiCA licensing · tax reports
status draft · waiting for your review
sample run · demo data
01 / The problem it removes
The page keeps slipping to next sprint.
You know which page would win the question. Writing it keeps losing to product work, and a competitor fills the gap.
Without it
- The prompt gap is known, but the page never gets written.
- Generic AI content reads generic, and engines skip it.
- Nobody wrote down which question the page must answer.
With the Articles agent
- Every draft starts from a tracked prompt you are losing.
- Drafts carry your real facts, comparisons and citations.
- Sections are structured so answer engines can lift and cite them.
02 / How it runs
A loop, not a task list.
- 01
Pick
A losing prompt with demand becomes the target.
- 02
Draft
A full article with sources, tables and FAQ blocks.
- 03
Review
Read the draft in the drawer, then download or publish it; nothing publishes itself.
- 04
Measure
Citations for the target prompt are tracked after publish.
03 / What it can do
What the agent does
Evidence-based
Your real fees, features and certifications, cited.
Rubric gate
Humanization, evidence, citations, structure, voice, language.
Comparison pages
vs-pages built from competitor evidence.
Versioned drafts
Append-only versions; Publish is explicit.
Post-publish tracking
Whether the engines started citing it.
04 / In the product
A draft lands. You review it, then publish.
Every Articles agent draft arrives with its keyword, outline and a Markdown body that cites at least three sources; a draft that fails the rubric is never saved. Versions are append-only, and Publish is a separate, explicit step that points public delivery at the version you chose.
target "coinbase alternative for spanish users"
length 2,400 words · 6 sections · FAQ block
includes fee table · MiCA licensing · tax reports
status draft · waiting for your review
sample run · demo data
05 / Questions
Before you install it.
The full Markdown body with its outline, the primary and secondary keywords, the target length and the reasoning behind it, plus the evidence references and citations it answers. You choose the format per workspace: brief, standard or deep dive.
A persisted Design Guide with a personality marker and a Product information document; generation is refused without them. The queue itself comes from your content strategy document, and choosing Generate now shows the title, keyword, target length and the credit quote before anything runs.
An executable rubric scores every version on humanization, evidence, citations, structure, brand voice and language, each with its own threshold; language must score 100. A draft also needs at least four themed sections, three evidence references and three citations. Promotional phrasing is a hard fail.
Lite and Pro. Lite is $129 a month, or $1,290 billed annually, and grants 2,000 credits each month; a workspace without the entitlement opens the plan picker instead of generating. The free plan covers the GEO, SEO and X agents.
Ships on the Lite plan
Start with the free analysis.
Free analysis, 5 credits that never expire, no credit card. The free analysis previews the Articles agent; it briefs itself on your site once you are on Lite, $129 a month.