The machine that processes your files has its outbound network access blocked by default, and that block is checked continuously — not asserted once in a policy document and never verified again.
Why outbound access is the real risk, not just inbound
Most security conversations focus on keeping attackers out. For a processing pipeline handling other people's confidential files, the more relevant question is what happens if something inside gets compromised or misbehaves — a vulnerable dependency, a bad update, a bug. If that machine can freely reach the internet, your data can leave along with whatever else goes wrong.
What's actually in place
The processing machine's route out is blocked at the network level, with one cable running only to the machine that dispatches jobs to it — every other path out is closed, and that closure is checked on an ongoing basis, not a setting configured once and assumed to hold.
Where this breaks down — said plainly
- This blocks network exfiltration. It's one control among several (encryption, malware scanning, retention limits), not a claim that it alone makes every other safeguard unnecessary.
- "Checked continuously" means a real recurring check, not an absolute real-time guarantee — there's a window between checks, by the nature of any periodic monitoring system.
Related guides
Where this fits on the site: this is part of how every job runs here, across all services. See Services & Pricing for what's actually sold, or go straight to Choose your job.