What SEO Crystal Ball checks
SEO Crystal Ball reviews common page and site signals, including metadata, headings, content clarity, images, internal links, structured data, indexability, mobile basics, performance clues, and local SEO signals. A focused page scan inspects one URL; bounded same-site scans can reveal repeated patterns across a larger group.
The goal is not to overwhelm you with every possible technical omen. It is to separate verified problems from review prompts, show which findings deserve attention, and connect them to a practical next step.
How to run a scan
- Open SEO Crystal Ball from the tools page.
- Enter a public URL or choose the available site-scan options.
- Set any page limit, discovery method, or additional URLs you need.
- Run the scan, then review the executive summary first.
- Open the priority findings, category scorecards, and affected pages.
Start with a focused page when you are diagnosing one landing page. Use a bounded crawl when you need to find repeated template issues, compare important sections, or inspect URLs discovered from the site or sitemap.
How to read the summary
The summary is the map, not the whole expedition. Review critical and high-priority findings first, then confirm the evidence on the affected page. A noindex directive, failed request, or broken canonical path deserves different treatment from a low-priority preview suggestion.
The score is an internal QA signal, not a Google ranking score. Read the category scorecards, severity, confidence, and supporting evidence together. A modest score with clear, verified fixes is more useful than a perfect number built on a partial response.
How to run the AI Citation Monitor
- Open AI Citation Monitor (Experimental) after signing in with a paid entitlement.
- Enter one neutral customer-style research question and optional location context.
- Add the target brand, domain, assigned URLs, and any competitor domains you want the parser to recognize.
- Select OpenAI, Perplexity, or both, then run the probe.
- Review cited and consulted sources, brand mentions, citation share, first citation position, and competitor appearances.
- Compare only with a prior probe that uses the same query and provider context.
The monitor is separate from the scored audit. It records evidence from grounded provider APIs and does not reproduce every logged-in ChatGPT or Perplexity experience. Read How to Read AI Citation Monitor Results before treating a one-off probe as a trend.
Provider setup, billing, and privacy
The monitor uses optional bring-your-own OpenAI and Perplexity API credentials. Consumer subscriptions do not include API billing. Provider keys are encrypted in the account-scoped secret store and are not placed in scan requests, reports, browser storage, or logs.
The current prototype limits provider calls per account and saves answer history for dated comparisons. Queries and location context are processed by the selected provider. Review Data Providers, API Keys, and Monitor Privacy for current setup, retention, deletion, and third-party-processing details.
How to move from report to fix
| Finding type | Start with this guide |
|---|---|
| Missing title or meta description | Metadata |
| Missing H1 or thin content | Headings and Content |
| Missing alt text or weak links | Images and Links |
| Noindex, canonical, robots, sitemap, JSON-LD | Schema and Indexability |
| Many mixed warnings | How to Prioritize SEO Fixes |
How to apply fixes on your platform
Use the finding category and affected-page list to decide where to look. Metadata and headings often live in the page editor or SEO settings. Image text lives in media settings. Repeated indexability, schema, or layout findings may come from a plugin, app, shared template, server rule, or custom code.
On static or custom sites, fixes may require editing HTML, discovery files, server rules, or JSON-LD directly. Preserve a working version, trace repeated findings to their shared source, make one logical change at a time, and verify the public response after deployment.
When to run another scan
Re-scan after a meaningful group of changes has reached the public site. Confirm the expected HTTP response first, then compare the new result with the earlier scan. Browser caches, edge caches, and search indexes update on different schedules, so a clean audit does not guarantee an immediate change in search results.
Turn the Report Into a Shorter, Better List
Run SEO Crystal Ball, verify the strongest findings, and use the linked fix guides to move from evidence to action without chasing every warning.
Run a Scan