Audits (Beta)

What the audit checks

A site audit starts from a website URL, discovers same-site pages through sitemaps, and processes them in bounded batches. It evaluates technical SEO, machine-readable content, structured data, AI bot access, trust and security signals, and other AI-readiness evidence.

• Sitemap discovery determines the site-wide page inventory.
• Page checks record pass, warning, failure, severity, evidence, and a recommendation where applicable.
• The homepage receives additional screenshot and performance enrichment.
• Reports aggregate page-level findings while keeping the underlying audited page available.
• The headline Site score uses the latest finished whole-site run. Pages checked is successfully processed pages out of accepted sitemap URLs. Open issues counts issue rows in that run's report snapshot, while the detail states how many are critical or high severity. Whole-site and single-page history tables each show at most the 25 most recent rows.

Run and monitor an audit

1. Confirm the canonical site URL Use the public website origin you want to audit, including the correct scheme and host.
2. Start the audit PromptSpark.ai discovers sitemap pages before creating and processing bounded batches.
3. Watch status and failures A large site can take time. Review discovery, batch, and processing states instead of restarting immediately.
4. Open page evidence Use category, severity, and status filters to move from the executive score to the exact affected page and check.

Prioritize fixes

Start with blockers that affect many important pages, such as inaccessible sitemaps, AI bot restrictions, broken canonical or structured-data patterns, and content that cannot be read without client-side execution. Then address page-specific quality findings.

Re-run after deployment and compare the exact checks. A generated fix is a draft; validate it in the site's framework, design system, and release process before publishing.

{"Site score is a weighted average of applicable category scores" => "Technical SEO 25%, Structured Data 20%, AI Readiness 25%, Trust and Security 15%, and Performance 15%. Within a category, passed checks score 100; warnings score 40, 55, 70, or 85 from critical to low severity; failures score 0, 15, 30, or 45. Not-applicable checks are excluded, and remaining weights are normalized before rounding to a whole number."}

Read an audit finding in context

Status and severity help sort work, but the evidence determines what actually needs to change. Open the affected page and check, then compare the recorded evidence with the live implementation. A site-wide template issue deserves a different response from an isolated content warning.

1. Confirm applicability Make sure the check applies to the page type and intended experience. Some checks can be not applicable by design.
2. Identify the scope Determine whether the finding affects one URL, a template, a subdirectory, or the whole site.
3. Find the root cause Group findings created by the same CMS field, component, rendering rule, redirect, or infrastructure setting.
4. Validate the proposed fix Test accessibility, SEO, structured data, security, design, and performance implications before release.
5. Re-run and compare Confirm the exact check and page changed after deployment, while preserving evidence of any remaining exceptions.