Answers for the moment something does not add up.

Yes. A new account can begin on the free plan and run a compact audit. The current Plans page shows the exact site, crawl, and URL allowances available to the workspace.

Use the final HTTPS homepage on the canonical hostname. Leave out tracking parameters, fragments, and temporary redirects. Confirm that the normalized hostname is the property you intend to audit.

No. Every step can be skipped, the full route can be skipped, and skipped work can be restored later. Progress follows real workspace records, so another teammate can complete a step elsewhere in the app.

Yes. They are separate connections and can use different identities. Authorize with the account that can access the exact property needed for each source.

Authorization confirms the identity, not property access. Check that the chosen Google account is a user on the required property and that the hostname, protocol, or domain property matches the audited site.

Refresh timing depends on the source and workspace capability. Read the source timestamp shown in the report. The page can be newly opened while the underlying provider data is older.

It should not. Search Console, GA4, advertising, and customer-specific connections are separate. Reconnect only the affected record, then verify the selected account and property.

Standard balances coverage and speed. Fast is suited to a small release check. Deep explores more click levels with gentler concurrency. Reuse the same profile and scope when a fair comparison matters.

The fetch phase can finish before the evidence snapshot is written. Large crawls may stay in persistence for several minutes with no change to the fetched count. Wait for completion or a visible failure state before retrying.

The crawl stopped after reaching its configured or plan allowance. Reports describe the captured URLs, not the complete site. Use the crawl log and coverage indicators before drawing site-wide conclusions.

Regular and competitor audits respect robots.txt. A verified owner can use the deliberate control for an authorized test that ignores it. That option should never be used on a site you do not control.

The host, firewall, or CDN may have rate limited the crawler. Review the crawl log, test a sample URL, reduce concurrency, and allow the crawler when you control the infrastructure.

You can audit an ordinary public site when you are permitted to inspect it. The product blocks major platforms and other unsafe targets, respects robots rules, and does not provide owner-only controls for a competitor project.

Scope, parameters, rendering, robots behavior, redirect handling, and issue definitions vary. Compare repeatable trends within the same configuration before treating two totals as directly equivalent.

No. It is a Swiss Knife SEO summary of weighted issue density. Use it as a trend signal, then inspect severity, affected pages, and scope before deciding what the movement means.

Yes, if the later crawl captured fewer difficult pages or used different scope. Report score movement beside crawled URL count and configuration so the comparison remains honest.

Apply the relevant filters first, then use the report CSV export. For a full audit handoff, use the crawl-level spreadsheet or bundle. Confirm the saved crawl and segment before sharing.

It creates a Markdown version of the report or issue list with readable headings and rows. Review sensitive context before pasting it into any external assistant.

Measured project claims should link to supplied evidence. An uncited line should be treated as a suggestion, draft, or inference, then checked before it becomes implementation work.

It estimates internal authority while distinguishing contextual body links from repeated navigation, footer, breadcrumb, and component links. Read the score with inbound source and placement evidence.

It receives internal links, but none from contextual main content. It may still be discoverable, yet it lacks the editorial relationship that a relevant body link can provide.

No. Confirm topical relevance, reader value, placement, existing anchors, and destination intent. A smaller set of useful links is better than inserting every machine-generated opportunity.

Raw HTML is the first server response. Rendered HTML includes content created after browser scripts run. The comparison exposes important fields and links that depend on successful JavaScript.

No. It summarizes access and interpretation signals. A citation depends on the monitored question, engine, source selection, competing pages, and answer variability.

Generated answers are variable. Use repeated runs, the preserved answer history, sampling range, and stability label instead of treating one observation as permanent.

No. An answer can name a brand without linking to its site, and it can cite a page without naming the brand. Store and interpret the two signals separately.

Begin with real buyer questions, important page topics, and useful Search Console demand. Record market and language, and write enough context to reflect the decision the searcher is making.

Rank Tracker stores a controlled country, language, and device observation. Search Console aggregates real impressions and positions across many searches. Their methods and time windows are different.

No. It monitors known links that you add or import. Use it for links whose loss or change would justify action, not as a complete web-wide backlink index.

The source could not be checked reliably. A timeout, block, or server failure does not prove that the link disappeared. Retry later and inspect the final source response.

Provider-backed analysis, rank observations, AI visibility checks, writing, paraphrasing, and some advanced workflows can use metered services. Each tool's About panel and locked state should say whether a provider call can occur.

No. A saved report belongs to an existing crawl. A new allowance is used only when a new crawl is started manually or by schedule.

No. It may prefill safe input, but the recipient must explicitly choose Run or Refresh. A page load should never trigger a paid provider call.

Public detail pages can remain visible so teams can evaluate a deliverable. Execution follows the active plan and add-ons. The page should identify the required capability before any provider call.

Use the live Plans page and the active workspace notice. Packaging can change as the product grows, so old screenshots and support messages are not the source of truth.

An AI fix is a contextual remediation plan. A patch is implementation-ready markup, configuration, or code for a specific pattern. Both require stack review, testing, and rollback planning.

After the deployed result is confirmed by a later crawl, monitoring check, or direct evidence. Ready to verify is the correct state when implementation is live but confirmation has not arrived.

Yes, when the workspace supports that workflow. Add a reason, owner, and review condition so an intentional exception does not look like forgotten work.

A Viewer reads evidence and alters nothing. For comparison, billing and role assignment sit with an Owner, operational settings and approvals sit with an Admin, and the work itself gets built and launched by a Member. Pick the least permissive role that still lets the person finish the job.

Use a protected deployment secret or credential store. Send the token in the Authorization header, never in a URL, issue description, screenshot, or shared document.

Use the reset link on the sign-in page and follow the email promptly. If the message does not arrive, check the address, spam folder, and whether the account signs in through a different identity.

Opening a saved thread intentionally restores its messages, site, and artifacts. Use New task when the new investigation should begin with a blank composer.

No. It is scoped to supported work around the selected site's audit, search, content, accessibility, analytics, advertising, and conversion evidence. Rephrase unsupported work around a valid product outcome.

No. Site evidence is an explicit choice. No site context keeps the run brief-only, while a selected project can add relevant audit and connected evidence.

No. It prepares analysis, drafts, exports, mappings, and monitoring views. Provider campaign publishing and editing remain separate reviewed actions.

Automation finds only some deterministic barriers. Keyboard operation, reading order, screen reader clarity, zoom, form recovery, and content meaning still need human evaluation.

No. It is a prioritization signal. A conformance statement requires evaluation against the applicable criteria, including tests that automation cannot complete.

Include the exact page, workspace site, local time with timezone, browser, device class, action, visible message, retry result, and a screenshot when useful. Remove credentials and private customer data.

Name the source, date range, site, expected value, observed value, and the report where it appears. Compare the source timestamp and scope before sending the case.

Use the app help search, press Ctrl and K to find a report or tool, or search this help center by task, feature, error phrase, or SEO term.