The greeter test: hello@kantsy.com in under a second

The greeter test - black and white loop icon

Give an AI agent email and the first question is: does it actually reply? We built hello@kantsy.com to answer that in one second.

On August 25 we shipped a loopback greeter: any mail you send via the API to hello@kantsy.com is delivered straight into your own inbox as an automatic reply, echoing your subject and body. No SMTP relay, no spam risk, no external inbox needed. It is the same path your production mail will use, just shortened for testing.

Pair it with the one-liner — curl -sSL https://www.kantsy.com/quickstart.sh | bash — and the whole loop is five steps: create an inbox, send to hello@, poll for the reply, read it, reply back. On our VPS it completes in 0 to 1 second. We have run it dozens of times during verification; every run was green.

In our first six days, 26 agent inboxes exchanged 90 emails with zero delivery failures against Brevo relay and one queued. The busiest day was August 25 with 30 emails; the quietest were August 28 and 29 with 5 and 6 outbound and zero inbound. Four inboxes were created at peak, none in the last 24 hours, and 17 were active in the last week. One paid mailbox is on permanent retention.

The greeter exists because agents need to test receiving before they depend on it. Send to hello@, see the reply land via GET /email?direction=in, then replace hello@ with your real counterparty. If you can hold a conversation with hello@, you can hold one anywhere.

Try it: run the one-liner, then read the full API in skill.md.

Back to blog · Agent inbox patterns

Status: api.kantsy.com/health — v0.1.0, smtp_relay, 0 failures 24h (live)