e2a

e2a provides the open-source email infrastructure that applications and AI agents need to send transactional emails reliably. It eliminates the need for public URLs, port forwarding, or shared hosting platforms by running its own mail server on a single domain. The service evaluates SPF, DKIM, and DMARC records before delivering messages, ensuring deliverability while maintaining full control over email content through an optional human-in-the-loop approval gate.

The architecture separates inbound identity verification from outbound delivery: incoming emails are evaluated against DNS-based authentication protocols to confirm sender legitimacy, then delivered via SMTP or HTTP. Outbound sends can be routed through upstream relays (such as SES) or sent directly to other agents using the same protocol stack. This separation allows applications to remain decoupled from email infrastructure concerns while still receiving authenticated inbound messages.

The API supports multiple delivery channels including REST polling, WebSocket streams, and MCP tool integration for agent frameworks like Mastra, OpenAI Agents SDK, and LangChain. Templates are rendered server-side at send time using template variables, enabling dynamic content without client-side rendering. Contact management includes account-level CRUD operations with bulk import capabilities, while scheduled sends can defer delivery up to 90 days in advance.

The hosted service offers a single-page dashboard for managing agents, messages, and approval workflows, along with an MCP server that enables agent frameworks to interact directly with the email infrastructure without additional tooling. Self-hosted deployments use Docker Compose to run PostgreSQL databases, S3-compatible storage (MinIO), and the full SMTP relay stack on a single machine or cluster.