Practical guides for agent email infrastructure.
Long-form guides explain the architecture, implementation sequence, security boundaries, and failure modes behind reliable email workflows for agents and tests.
The mails0 blog is an implementation library, not a news feed. Each guide answers one durable developer question, links to runnable documentation, states the operating tradeoffs, and connects the reader to the smallest next product action.
Build the full email loop
Start with the problem you are solving, then move into an implementation path for MCP, Playwright, Cloudflare, OTP retrieval, event delivery, security, or measurement.
What is agent email?
Learn what agent email is, why AI agents need dedicated inbox identities, how tools and permissions work, and when to use hosted or self-hosted infrastructure.
Open guide →Give an AI agent an email address
Create a scoped temporary inbox, connect it through CLI or MCP, receive a real message, retrieve a verification code, and upgrade the identity safely.
Open guide →Automate verification with Playwright
Build a reliable Playwright email verification test with an isolated inbox, explicit waits, OTP or magic-link validation, cleanup, and actionable failure evidence.
Open guide →Email MCP server guide
Understand email MCP server architecture, configure mails-agent-mcp, protect mailbox credentials, choose tools, and decide when HTTP or SDK automation is better.
Open guide →Self-host agent email on Cloudflare
Plan and deploy a self-hosted agent email API with Cloudflare Workers, D1, R2, Email Routing, custom domains, outbound delivery, security, and operations.
Open guide →Temporary vs permanent mailboxes
Choose between temporary, permanent, and self-hosted AI agent mailboxes using lifecycle, identity, security, sending, retention, reputation, and cost criteria.
Open guide →Secure email for coding agents
Protect coding-agent email with dedicated mailboxes, scoped credentials, local secret storage, receive-only bootstrap, link validation, approvals, and audit evidence.
Open guide →Agent email API checklist
Evaluate agent email APIs across identity, inbox retrieval, OTPs, search, threads, webhooks, SDKs, MCP, security, deliverability, hosting, and operations.
Open guide →Test OTP without mocking email
Test real OTP email delivery with isolated inboxes, bounded waits, multilingual code extraction, stale-code handling, CI isolation, and final verification assertions.
Open guide →Email infrastructure for autonomous agents
Design autonomous-agent email across identity, events, untrusted content, approvals, idempotency, retention, deliverability, monitoring, and emergency shutdown.
Open guide →Inbox polling vs webhooks
Choose between polling, long waits, webhooks, and event streams for agent email based on latency, reliability, retries, concurrency, security, and cost.
Open guide →Email prompt injection defense
Defend AI agents against prompt injection in email bodies, links, attachments, threads, and headers with scoped retrieval, policy checks, and approvals.
Open guide →Measure agent email conversion
Measure agent email growth from qualified visit to mailbox bootstrap, first inbound message, completed workflow, durable claim, and retained authorized use.
Open guide →Build vs buy agent email
Decide whether to self-host, use a managed agent inbox, compose cloud primitives, or combine an open mailbox layer with an email delivery provider.
Open guide →Turn the guide into one real test.
Use a temporary mailbox to prove the first inbound event, then keep the article open while you harden credentials, retention, and deployment.