All documentation

Optimize

Page issue queue

Prioritize known owned pages using saved gaps, draft status, diagnostic findings, linked evidence and recorded outcomes.

Open Actions → Page issues to bring known owned pages, saved citation gaps and optimization work into one place. Start with a page, review the evidence, then open its existing workflow to decide what to implement.

The page issue queue for the fictional Quillstone project: 13 known pages, all 13 with actionable work, and its comparison page with three actionable issues, the first a Site audit item saying 16 SEO rules flag the page

Which pages appear

The queue groups URLs from the current project's prompt-to-page mappings, citation gaps with target URLs, optimization opportunities, saved diagnostic URLs and the project's newest completed site audit. Paused prompts retain their saved mappings. URLs use the same normalization as the page store: tracking parameters, fragments and trailing slashes are removed; meaningful query parameters and distinct hosts remain. Pages must belong to the project's registrable domain. Unsafe links and off-domain records are excluded and counted (api/services/page_issues.py#inventory_statement).

This is known inventory, not a complete site crawl. Pages seen only in Google integrations or the global URL store do not enter this first version. “No saved issues” means there is no saved gap or optimization for that page; it is not a health check or a clean audit result. Gaps without a target URL have their own count and link to Citation gaps.

Only the project's newest completed site audit feeds the queue. A running or failed audit is skipped when choosing that audit, so it never hides the previous completed audit's results while a new audit is in progress or after one fails; a stopped audit that completed counts (api/services/page_issues.py#inventory_statement). Each URL that audit fetched, could not fetch or found blocked by robots.txt appears as one Site audit item, labelled with a "Site audit" badge and linking to that URL's report in the site audit.

The screen reads up to 2,000 newest records from each of the five sources. If a source exceeds that limit, a warning names it. Counts, filters and ranking describe the inspected records, including when the inventory is limited.

Choose the next page to work on

The default order is actionable issue count, then known prompt count, then URL. An issue is actionable when it needs action or has a draft ready. You can instead sort by known prompts or URL, search URL text, and filter by work status. Results show 25 pages at a time; summary counts describe inventory before your filters.

Known prompt counts combine mappings with up to 20 prompt examples per issue. They are not complete prompt-reach or visibility metrics. Each issue shows its original citation-gap impact or optimization score and available effort label. Those two scores are not combined into a new AEO score or predicted lift.

When a citation gap already has a linked optimization for the same normalized page, it appears once under that optimization. Different target pages remain separate. Expand linked prompts and up to five cited-source examples to inspect the context, then choose Review evidence and next steps to open the exact saved issue, even if it is outside the normal list's first 100 records.

Follow the work

  • Needs action: a saved gap or opportunity has no displayed ready/applied draft.
  • Draft ready: an open optimization recommendation is available for review.
  • Applied: a recommendation was marked applied, or a citation gap is in an implemented, monitoring or recorded-outcome state.
  • Stale analysis: the parent analysis, displayed draft or owned-page mapping changed. Review and refresh the existing workflow before using old advice.
  • Dismissed: the parent issue was dismissed.

Optimization display follows the existing lifecycle: a non-dismissed applied recommendation still inside its measurement window takes precedence over newer drafts; otherwise the newest non-dismissed recommendation appears. The queue does not list every historical recommendation. Stale or dismissed parent state takes precedence over a ready draft. An applied date and outcome, when present, are saved observations; they do not prove the edit caused a performance change.

Use Optimizations to review drafts, approve handoff, mark changes applied and inspect measurements. Opening the queue only reads saved records. It does not run generation or refresh evidence, and the platform does not modify your website.

Access and limits

Company members, editors and viewers on every plan can read their project's saved queue. Linked generation, review, export and measurement actions keep their existing role and plan requirements. There is no new paid gate, integration or collection job. Saved page diagnostics adds explicit policy, HTTP, HTML, structure, response and browser-render checks, saved evidence and recheck history. Each exact URL's latest completed diagnostic appears as a separately labelled bundle with no impact score. Several exact diagnostic URLs can share one normalized queue page. Running or interrupted rechecks retain the previous completed queue result.

A bundle with review findings is Needs action; otherwise it is Diagnostics unknown if any check is unknown, or Diagnostic checks passed. A first run without completed evidence is unknown. A passed bundle is not a complete page health verdict. Opening the queue only reads; company members or editors with the existing robots diagnostics entitlement can open a page's diagnostic history and start an explicit recheck. Saved diagnostic evidence is readable by every current member. Rendering, readability and response timing are single lab heuristics, not field measurements. Third-party crawler access and field (Core Web Vitals) performance remain outside these checks.

A Site audit item is Needs action when any SEO rule flags the page in that audit, and says how many rules do. Informational rules are not counted, so the number can be lower than the URL report's Needs review count, which includes them (api/modules/website_crawler/reporting.py#refresh_url_issue_counts). A page no counted rule flags is Diagnostics unknown, never passed: rules the audit could not assess are not passes. A page the audit could not read, or that robots.txt blocked for DiscoveredByBot, is Diagnostics unknown too, with that reason (api/services/page_issues.py#page_issues). The "Site audit" badge is what tells it apart from a saved page diagnostic bundle. Open audit evidence opens the page's URL report, which every current member can read; starting a new audit keeps Site audit's own company member or editor and plan requirements.

Last verified 2026-10-15

Start monitoring your AI visibility.

See how AI search engines talk about your brand.

Free to start. No credit card required.