Full articles, not a paragraph generator: real research, a quality score with an automatic rewrite loop, and a house style built to strip out the AI tells other tools leave behind.
3 free articles · Grounded in your real data · Cancel anytime
Definition
AI content generation in inkSTR is the process of turning a researched keyword into a complete, publish-ready article: title, meta tags, structured body, FAQ section, images, and internal links, written in your brand voice from your real business data. It runs on the Claude model family with a fixed editorial framework, then grades and repairs its own output before anything reaches your calendar.
Key takeaways
Included in every plan. Last reviewed .
How it works
Every article starts from real keyword and competitor research, not a blank prompt. The system enforces no-fabrication rules, brand voice, and E-E-A-T structure, then scores the result and automatically rewrites anything that falls short.
The engine pulls the live SERP for the target keyword, reads the top-ranking pages, and collects real questions people ask about the topic.
Property or product details, Knowledge Hub entries, verified statistics, and writing guidelines are assembled into the brief so the article can be specific.
The article follows a fixed structure: answer-first opening, key takeaways, question-based sections, FAQ, and a call to action that fits your business model.
The draft is scored, rewritten if it falls short, scanned for AI tells and unverifiable claims, then finished with images, links, and schema.
What you get
Grounded in the real pipeline, not a marketing checklist.
Every article draws on real keyword data, competitor analysis, and your own business details, so it says something specific instead of generic filler.
Articles are scored on a 10-point rubric after generation. Anything below an 8 triggers an automatic rewrite before it ever reaches your calendar.
A layered detection system, banned phrases, sentence-pattern rules, and statistical prose analysis, catches and rewrites the corrective-negation, rule-of-three, and filler-intensifier patterns that read as obviously AI-written.
Every article gets AI-generated or library-matched images, with alt text and captions handled for you, and video enrichment pulled in from YouTube or Vimeo when it genuinely fits.
Side by side
Where the difference shows up in practice. No invented numbers on either side.
| Where it matters | InkSTR | A generic AI writer | Why it matters |
|---|---|---|---|
| Starting point | Live keyword and competitor research plus your stored business facts. | Whatever you type into the prompt box. | Specific inputs are what let an article say something a reader cannot get from any other site. |
| Quality control | Scored on a rubric, rewritten below 8/10, then checked by deterministic detectors. | You read the draft and decide. | A gate that runs on every article catches the bad ones at 2 a.m. when nobody is reading. |
| Statistics and sources | Only figures from your verified stats library or approved industry data; invented source names are caught and removed. | Plausible numbers with no provenance. | A fabricated statistic under your brand name is a trust problem that outlives the article. |
| Finished output | Images, alt text, captions, internal links, FAQ schema, and meta tags included. | A block of text to format and illustrate yourself. | Publish-ready means the article can go live on a schedule without a human in the loop. |
Starting point
InkSTR: Live keyword and competitor research plus your stored business facts.
A generic AI writer: Whatever you type into the prompt box.
Why it matters: Specific inputs are what let an article say something a reader cannot get from any other site.
Quality control
InkSTR: Scored on a rubric, rewritten below 8/10, then checked by deterministic detectors.
A generic AI writer: You read the draft and decide.
Why it matters: A gate that runs on every article catches the bad ones at 2 a.m. when nobody is reading.
Statistics and sources
InkSTR: Only figures from your verified stats library or approved industry data; invented source names are caught and removed.
A generic AI writer: Plausible numbers with no provenance.
Why it matters: A fabricated statistic under your brand name is a trust problem that outlives the article.
Finished output
InkSTR: Images, alt text, captions, internal links, FAQ schema, and meta tags included.
A generic AI writer: A block of text to format and illustrate yourself.
Why it matters: Publish-ready means the article can go live on a schedule without a human in the loop.
Who it is for
The same engine adapts its audience, vocabulary, and calls to action to the business you run.
Guest-facing guides that name your actual properties, link to your booking pages, and pull real photos from your image library.
ExploreOwner-facing articles about fees, regulations, and revenue that position your company as the only management option named on the page.
ExplorePlain-English service explainers for each town you cover, written for homeowners and free of the jargon and invented claims that hurt local trust.
ExploreMeasured results
Every figure below is measured in Google Search Console, Google Analytics, or the business's own records, never estimated by a third-party tool.
788 → 5,195
Monthly Google Search clicks in 11 months
Hemlock Hills Cabin Rentals
Every article on the site was generated through this pipeline, grounded in the brand's real cabins and booking pages. Traffic figures are measured in Google Search Console, not estimated.
Read the Hemlock Hills Cabin Rentals case studyYour actual business. It’s grounded in your real property, product, or service details and strict no-fabrication rules across every generation path, never invented statistics, ratings, or claims.
30 articles per month per project on Grow and Scale, with Dominate removing the cap. Each project runs its own independent content calendar, and every plan starts with 3 free articles.
Yes. Every article opens in a visual editor, and you can regenerate, edit, and review before it publishes, or let it auto-publish on schedule, your choice per project.
Articles are written with Anthropic's Claude models. The model is only one part of the system: the research brief, the editorial framework, the scoring gate, and the post-generation detectors are what make the output publishable, and they run on every article regardless of model version.
Most articles finish in a few minutes. Research-heavy topics for established sites can take 15 to 20 minutes because the engine reads more competitor pages and runs more repair passes. Generation runs in the background, so you can leave the page.
The house style bans the phrases and sentence rhythms readers now recognize as machine-written, and a statistical pass measures rhythm, filler intensifiers, and repeated patterns across the whole article. Drafts that trip those checks are rewritten and re-checked before they are kept.
Yes. The engine adapts its audience, seed vocabulary, and calls to action to the business model you set up: short-term rental, property management, SaaS, home services, real estate, and custom local businesses each get their own rules.
Go deeper
Guides, playbooks, and tools that explain the ideas behind this feature.
A few more capabilities worth a look.
A content calendar built from real search volume, not a keyword list
ExploreGet cited by ChatGPT, Perplexity, and Google AI Overviews, not just ranked
ExploreEvery new article links your old ones, without a manual audit
ExplorePoint inkSTR at your site and see the content strategy it builds. Your first 3 articles are free.
3 free articles · Grounded in your real data · Cancel anytime