Features
Every check, and what each one is for
RankInAI runs roughly seventy deterministic checks across seven weighted categories. Below is what each category covers and the platform features around it.
Technical Accessibility
Can AI crawlers reach, load and index your pages at all?
What is checked
- HTTPS coverage and certificate validity
- HTTP status codes and redirect chain length
- noindex directives and page indexability
- robots.txt presence, rules and blanket blocks
- Named AI crawler access (GPTBot, PerplexityBot, ClaudeBot, Google-Extended and others)
- XML sitemap discovery and URL count
- llms.txt detection
- Canonical URL coverage and self-reference
- Mobile viewport and language declaration
- Broken internal link sampling
- Image alt-text coverage
- Crawl success rate and server response time
- Number of distinct discoverable pages
Entity Clarity
Can a model state who you are, what you do and who you serve — without guessing?
What is checked
- Business name prominence in title, H1 and opening copy
- A plain statement of what the business does
- Explicit definition of the customers served
- Service and product specificity across pages
- Geographic clarity: address, service area, areaServed
- About page substance
- Name consistency across page titles
- Title tag quality and length
- Meta description coverage
- Duplicate title and description detection
- Open Graph and Twitter metadata
Content Authority
Is there enough original, specific, expert content to be treated as a source?
What is checked
- Average content depth per page
- Thin and near-empty page detection
- Presence of explanatory or educational content
- Concrete facts and figures versus qualitative claims
- Empty marketing filler detection
- Unsupported superlative detection
- Content freshness and visible dates
- Internal linking depth
- Heading hierarchy and single-H1 structure
- First-party evidence: case studies, original data, customer proof
- Duplicated body copy across templated pages
- Homepage explanatory substance
Answer Readiness
How easily can a single passage be lifted and used as a direct answer?
What is checked
- Question-shaped headings
- FAQ coverage and FAQPage markup
- Passage sizing — extractable paragraphs
- Lists and tables for structured information
- Explicit term definitions
- Comparison and alternatives content
- Front-loaded answers before elaboration
- Terminology alignment with customer language
- A quotable homepage summary sentence
- Process and how-to content
Structured Data
Is there valid, reasonably complete schema markup — not just a token block?
What is checked
- JSON-LD presence across pages
- Syntactic validity of every block
- Organization or LocalBusiness declaration
- Organization property completeness (logo, description, address, telephone, sameAs)
- sameAs profile links
- WebSite and BreadcrumbList schema
- Content-type schema: Article, BlogPosting, Service, Product, HowTo
- FAQPage markup on FAQ content
- Person and author schema
- Review and AggregateRating markup
- Agreement between schema name and the visible brand
Trust and Evidence
Is there enough verifiable substance to treat the content as reliable?
What is checked
- Published phone, email and postal address
- A dedicated contact page
- Named author attribution on content
- Stated credentials, licenses and expertise
- Outbound citations to authoritative sources
- Source transparency inside the prose
- Customer proof: case studies, testimonials, review markup
- Business legitimacy markers: privacy policy, terms, license number, founding date
- Clear attribution of content to the business
- Publication and update-date disclosure
Competitive Visibility
How do your observable signals compare with the competitors you supply?
What is checked
- Overall homepage signal comparison
- Structured-data coverage versus competitors
- Content depth versus competitors
- Answer-ready structure versus competitors
- Schema types competitors declare that you do not
Platform
Built to be used, not just looked at
The audit is the product, but the surrounding platform is what makes it usable week after week.
Deterministic scoring
The same website content always produces the same scores. No randomization, no black box — every point traces to a recorded observation.
Evidence panel on every finding
Each finding shows the exact values found: which URLs, which properties, which counts. You can verify any claim yourself in a browser.
Safe, polite crawling
The crawler identifies itself, honors robots.txt, never submits forms, never authenticates and never executes downloaded files. Private networks are blocked at the socket level.
Branded PDF reports
A full PDF with cover page, category scorecards, findings, action plan, page-by-page observations and methodology — with page numbers and clean page breaks.
Historical score tracking
Re-run an audit after making changes and see the score movement against your previous audit of the same domain. Included on Growth and Agency.
White-label reports
Agency plans replace the RankInAI cover branding with a client name and logo, keeping a small "Powered by RankInAI" line in the footer.
CSV export
Export findings and recommendations for your own reporting, tracking spreadsheets or client deliverables. Growth and Agency plans.
Team-ready architecture
The data model separates accounts, subscriptions, audits and reports so multi-seat access can be added without migrating your history.
Optional enhancements
Two optional layers, clearly separated
The core audit works fully without either. When they are enabled, the report labels their output explicitly so you always know where a statement came from.
AI narrative enhancement
When an OpenAI key is configured, a model rewrites executive summaries and explanations into clearer prose. It receives only already-computed scores and check outcomes — it never sees raw page content and can never change a score, a status or a piece of evidence. Enhanced text is marked AI-enhanced in the report.
Public-web search observations
When a search provider is configured, RankInAI runs a small fixed set of queries (brand plus service, brand plus location, brand reviews) and reports what is publicly visible. These are labeled public-web observations and are never presented as evidence of what any particular AI model will say.
See the checks running on a real website
The free homepage preview runs twelve of these checks on your own site without an account.
One-time audits from $49. Monthly plans from $29. All prices in US dollars.