A synced scrape is the shared copy: one crawl for the whole team, and the connection point for MCP clients.
Every scrape starts and stays on the machine that ran it. That is the right default for client data: the scrape itself is not uploaded until you decide a workspace should travel.
Syncing is the deliberate step that turns one person's scrape into a shared copy. It is also what makes MCP access possible at all.
A synced scrape is what Claude, ChatGPT, Cursor, and Codex connect to. They query the pages, issues, links, and groups in that run, not a spreadsheet you rebuilt for the chat.
Nothing reaches a client until you push it there. Re-sync after a fresh scrape and that is the context those tools read next.
Crawl evidence normally lives on one laptop, which means the person who ran it becomes the bottleneck for every question about that site.
A synced workspace lets anyone on the team open a teammate's scrape, run their own analysis, and build deliverables from the same pages instead of scraping the site again.
You choose which scrapes leave the machine. Sync is per workspace, not a blanket upload of everything on disk.
Unsync deletes the cloud copy only. The local crawl stays put, and MCP clients lose access to that workspace until you push it again.
Everything runs in one workspace, from the first crawl to the fix that ships.
Turn websites into SEO knowledge bases. Free to start.