Each refresh becomes a new version beside the last. Silkra compares the two field by field, so you read what moved instead of hunting for it.
Run it again from the version menu. The new pass reuses the last settings and lands as v2, v3, and so on.
Silkra lines up each URL against the run before it, from status and canonical to titles and body copy.
Added, removed, and changed pages land in one log, with regressions to status and indexing sorted first.
Every refresh lands in one log, ordered by impact. Status and indexing changes sit at the top, so you see what broke before you read what moved.
The summary groups a refresh by impact: regressions, new and removed pages, content, then titles and meta. Each page says what moved at a glance.
Pick a field and the log becomes a table: every page where it changed, before beside after. A sitewide title rewrite reads in one pass.
Refresh any site you track, confirm the fixes you approved actually went live, and ask the agent what a refresh means before you dig in.
Crawl a competitor in its own workspace and refresh it when you want to check in. New pages, removed pages, and rewritten copy land in the same log.
Approved fixes are checked against the next crawl, page by page. Each reads Implemented, Not updated, or Changed differently, so nothing ships on faith.
The agent reads the change log alongside the rest of the crawl, so it can say which edits touch indexing or internal links and which are just copy.
Most scraping tools work once. You crawl, export, and move on. When you come back, you run it again and compare two spreadsheets by hand, guessing at what moved in between.
Code solved this long ago with diffs. Websites need the same thing. In Silkra, every refresh is a new version and the change log is the diff: what changed, how it affects whether pages are fetched and indexed, and what to do about it.
Re-running a full crawl is slow, and heavy on the site being crawled. So a refresh starts by asking the server whether each page has changed since last time, using the same ETag and Last-Modified headers browsers rely on for caching.
When the server answers "not modified", Silkra carries the saved copy forward and moves on. Only pages that report a change, or cannot say, get scraped again. Refreshes finish sooner and put far less load on the site.
Everything runs in one workspace, from the first crawl to the fix that ships.
Crawl a site, then ask what needs attention.
Free to start. No credit card needed.
Jira
Linear