Staging builds, pages behind a login, client sites you do not control. Supply the facts you have and get a severity ranked plan. For anything public, the live audit measures it properly instead.
Output
Your brief, triage, or architecture recommendation will appear here.
The output is designed to be copied into a planning document and reviewed by a person before implementation.
Status, canonical, robots directives and whether anything is stopping the page being indexed at all.
Whether the page answers one query well, and whether the structure matches how that query gets answered.
Where the numbers you supply put you against the thresholds, and which of them is actually costing you.
Access problems are always first. A page carrying a noindex directive cannot be improved into ranking, so the content work behind it is wasted until that is fixed. The plan reflects that ordering rather than listing everything as equally urgent.
The free audit fetches your page and measures it. This one reasons about facts you supply, which is what you need when the page cannot be fetched: behind a login, on staging, not published yet, or on a client site you do not control. Use the audit when the URL is public. Use this when it is not.
No, and it says so throughout. It reasons only from what you type in. Anything claiming to have crawled a page it was never given access to is guessing.
Title, H1, meta description, whether the page is indexable, whether the canonical is correct, and your LCP, INP and CLS if you have them. Leave the ones you do not know blank rather than guessing, because a wrong number produces confident wrong advice.
Search Console gives you field data from real users, which is the number that counts. PageSpeed Insights gives you both field and lab. Lab data from your own laptop is the least useful of the three.
Yes, ranked by severity. Access problems outrank content problems, because a page that cannot be indexed cannot be improved into ranking.