Agent Mail Gateway

Secure email send/receive for AI agents, with human-in-the-loop approval

Problem

A developer built an IMAP+SMTP server for agents, signaling that agents need independent email identities for workflows. But granting agents email access is risky—mis-sends, phishing, or spam—and there's no approval or content-filtering layer tailored for agents.

Solution

A lightweight email gateway that exposes standard IMAP/SMTP to agents but enforces a rules engine and human approval before sending (one-click approve via Telegram or desktop notification). It also scans inbound links and attachments and keeps full audit logs.

Users

Developers building email-processing agents, automation workflow teams, and small businesses needing compliant approval

MVP

Agent-specific SMTP/IMAP endpoints; hold sends pending approval via push; auto-allow whitelisted domains; basic inbound link and attachment risk scan; audit log view and export

Revenue

Per-mailbox subscription; free tier with limited monthly approvals; team plan adds audit export and SSO

Why now

Agents are taking on real workflows where email is the core interface, yet security and approval layers are nearly nonexistent. Indie developers can build fast with open-source mail libraries.