Public Beta
Safe email for
your AI agents
Give your agent its own email address — with built-in protection against data extraction and prompt injection.
Works with any agent
Claude Code
OpenClaw
Cursor
Windsurf
LangChain
CrewAI
Security that scales with the stakes
L1
Receive Only
← [email protected]✓ clean
"Ignore prior instructions..."✕ blocked
Injection scanExfiltration defenseSender filter
L2
AI-Protected
AI scan active
Scanning inbound message
✓ No injection✓ Sanitized
AI security scanContent sanitizationInjection defense
L4
Full Lockdown
Refund: $450.00pending
Approve
Reject
Human approvalDraft reviewAuto-timeout
Built for agents, not admins
Terminal
$ sam inbox create --slug ops-agent --level L2
$ sam policy add ops-agent --allow-from "acme.co, partner.io"
✓ Sender allowlist active
$ sam send ops-agent --to [email protected] \
--subject "Your refund has been processed"
✓ Sent — policy checks passed, no PII detected.
Ready to get started?
Free to start. No credit card required.