What the desktop spider is best at
Ad-hoc forensic work. Custom extraction with XPath, crawling a staging site behind basic auth, comparing a JavaScript-rendered crawl against raw HTML, and building redirect maps for a migration. If that is your job this week, it is hard to beat.
Where it stops
No scheduling without extra setup, no retained history, no alerting, and crawl speed and memory are capped by your laptop. On a large site you learn to babysit a crawl - and nobody notices a regression that lands while nobody happens to be crawling.
What a cloud crawler adds
Scheduled recurring crawls, stored history so you can diff this week against last, alerts when new errors appear, shared access for a team or client, and issue tracking that remembers what you already fixed. That last part is what turns an audit into a workflow.
Crawl limits, honestly
Free desktop crawling is capped at a few hundred URLs; paid desktop licensing is unlimited but bounded by your hardware. Cloud tools price by pages crawled per month because crawling is the compute cost - SEO Smart Engine allocates monthly page budgets per plan and shows the remaining budget before you start a crawl.
A practical combination
Use the desktop spider for migrations and one-off extraction. Use a scheduled cloud crawl as the monitor that tells you something broke, with export to CSV, JSON or PDF so the findings live outside any single tool.