Agent inbox
A real, deliverable email address owned and operated by an AI agent rather than a person — used to send, receive, and act on mail programmatically.
An agent inbox is an email address that belongs to a software agent instead of a human. It has a real, deliverable address, it can send and receive mail, and the agent operates it through an API or MCP tools rather than a graphical mail client.
The point is autonomy: an agent with its own inbox can sign up for services, receive verification links, reply to customers, and catch alerts without a person relaying messages in and out.
Why agents need their own inbox
Most of the world still reaches people by email — confirmations, receipts, verification codes, replies. An agent without an inbox is cut off from all of it and depends on a human to forward messages.
Giving the agent its own address makes it a first-class participant: it can be written to, it can reply in-thread, and its actions are attributable to its identity.
How it differs from a shared mailbox
A shared team mailbox is built for humans clicking around a UI. An agent inbox is built for software: typed tools to list threads, read messages, reply, label, and set rules — with scoped keys controlling exactly what the agent may do.
FAQ
Is an agent inbox a real email address?
Yes. It's a deliverable address (like name@mailgent.dev) with proper DKIM and SPF, not a simulated one.
Can multiple agents share one inbox?
Each agent identity gets its own address and scoped key, so actions stay isolated and attributable.
Give your agent an inbox.
A real email address, a vault, 2FA, and an identity in one API call.