# Molted Email

> Molted Email gives each AI agent its own mailbox: it sends and receives email over a REST API, the molted CLI or a hosted MCP server. Inbound mail is screened for prompt injection before the agent reads it, every send is checked against guardrails the agent can tighten but not loosen, and a thread can be handed to a person.

Agents integrating with Molted should start with the skill references below: they cover sign-up, keys, sending, the inbox and every route.

## Agent references

- [API skill reference](https://molted.email/skill.md): every agent route, the policy rules and common patterns
- [CLI skill reference](https://molted.email/cli-skill.md): the molted CLI, command by command
- [Documentation](https://molted.email/docs): guides and reference for the API, CLI and MCP server
- [Documentation index for language models](https://molted.email/docs/llms.txt): every docs page as Markdown, in sidebar order (/docs/llms-full.txt has them all in one file)

## Product

- [Home](https://molted.email)
- [Pricing](https://molted.email/pricing): priced by mailbox: Free $0 (3 mailboxes, 3,000 sends a month), Solo $19/mo (10 mailboxes, 10,000 sends a month), Team $79/mo (50 mailboxes, 50,000 sends a month); more than 50 mailboxes, contact us
- [Pricing as Markdown](https://molted.email/pricing.md): every plan, limit, trial and overage rate as text
- [Features](https://molted.email/features)
- [Prompt injection protection](https://molted.email/features/prompt-injection-protection): Inbound mail is scored for injection before your agent reads it, and the messages that try to take over are quarantined.
- [Rules and guardrails](https://molted.email/features/rules-and-guardrails): Duplicates, cooldowns, suppressions and daily limits are checked on every send. Your agent can tighten them, never loosen them.
- [Inbox and handoff](https://molted.email/features/inbox-and-handoff): Threaded conversations your agent reads and answers, and a handoff to a person when a thread needs one.
- [Autonomy dial](https://molted.email/features/autonomy-dial): Every mailbox starts autonomous. Turn one down and first emails, or all of them, wait for your approval.
- [Multi-agent coordination](https://molted.email/features/multi-agent-coordination): Leases stop two agents emailing the same person at once, and every agent shares one set of checks.
- [Real-time events](https://molted.email/features/real-time-events): Deliveries, bounces and replies arrive as events, over a stream your agent holds open or signed webhooks.
- [MCP, CLI and REST](https://molted.email/features/developers): The same product over a hosted MCP server, the molted CLI and a REST API, for Claude Code, Cursor, Codex or your own code.
- [Humanizer](https://molted.email/features/humanizer): Off unless you turn it on. When it's on, a model rewrites each templated email in the tone you pick.
- [Agent-native email](https://molted.email/agent-native-email)
- [Molted vs AgentMail](https://molted.email/compare/agentmail): screening, duplicate sends, approvals, handoffs and pricing, side by side with sources

## Use cases

- [Use cases](https://molted.email/use-cases)
- [Support agent](https://molted.email/use-cases/support-agent): Answers your support inbox. Every message is screened before the agent reads it, and the hard ones go to a person.
- [Inbound lead agent](https://molted.email/use-cases/sales-agent): Answers people who wrote in or filled a form, asks the qualifying questions, and hands warm leads to a person.
- [Form to inbox](https://molted.email/use-cases/form-to-inbox): Website form submissions land in your agent's inbox as threads. It replies to the person who wrote and routes the rest.
- [Agent alerts](https://molted.email/use-cases/agent-alerts): An ops inbox for alerts, monitoring and vendor notices. Your agent triages it and escalates the ones that need a human.
- [Billing agent](https://molted.email/use-cases/billing-agent): Sends payment reminders to your customers once per invoice, stops following up when they reply, and hands billing questions to a person.
- [Customer success agent](https://molted.email/use-cases/customer-success-agent): Checks in with your customers at a pace they can take, and hands unhappy ones to your team.
- [Newsletter reply agent](https://molted.email/use-cases/marketing-agent): Handles the replies, unsubscribes and questions that come back to your newsletter and product updates.

## Security

- [Email prompt injection attacks](https://molted.email/security): ten attack categories, with examples and defenses
- [Instruction override](https://molted.email/security/instruction-override): "Ignore all previous instructions." The most direct attack, and it still works on unprotected agents.
- [Role play](https://molted.email/security/role-play): The attacker hands the model a new persona, one without the old rules.
- [System prompt mimicry](https://molted.email/security/system-prompt-mimicry): Fake [SYSTEM] tokens and control directives that read as if they came from the operator.
- [Delimiter abuse](https://molted.email/security/delimiter-abuse): Code fences, XML tags and separators that make injected text look like structure.
- [Data exfiltration](https://molted.email/security/data-exfiltration): Getting the agent to reveal its instructions, other mail, or the credentials it can see.
- [Encoding evasion](https://molted.email/security/encoding-evasion): Base64, hex and rot13. If the model can decode it, a payload can hide in it.
- [Context manipulation](https://molted.email/security/context-manipulation): Fake conversation history that tells the agent it is somewhere else, talking to someone else.
- [Multi-language attacks](https://molted.email/security/multi-language-attacks): Switching language mid-message to get past rules written for English.
- [Social engineering](https://molted.email/security/social-engineering): Urgency, authority and flattery. The tricks that work on people work on models too.
- [Token smuggling](https://molted.email/security/token-smuggling): Zero-width characters and look-alike letters: invisible to a reader, visible to the tokenizer.

## Free tools

- [Tools](https://molted.email/tools)
- [Prompt Injection Scanner](https://molted.email/tools/prompt-injection-scanner): Paste email body text, get a risk assessment across all ten pattern categories our production detector catches.
- [Email Template Linter](https://molted.email/tools/email-template-linter): Checks subject lines and HTML bodies for spam-trigger phrases, insecure URLs, missing unsubscribe links, and broken template variables.
- [SPF, DKIM and DMARC Checker](https://molted.email/tools/domain-checker): Enter a domain, get its SPF, DKIM, DMARC and MX status, problems ranked by severity, and the exact record to publish for each fix.
- [A/B Test Significance Calculator](https://molted.email/tools/ab-test-calculator): Plug in your control and variant numbers. Get p-values, confidence intervals, and conversion lift from a two-proportion z-test.

## Blog

- [Blog](https://molted.email/blog)
- [AI Email Automation: How Agents Send, Wait and Hand Off](https://molted.email/blog/ai-email-automation)
- [Human in the Loop AI Agents: Approval Gates That Work](https://molted.email/blog/human-in-the-loop-ai-agents)
- [AI Customer Support Agents: What They May Send](https://molted.email/blog/ai-customer-support-agents)
- [AI Email Security: How to Protect Agent Mailboxes](https://molted.email/blog/ai-email-security)
- [What's New: Journey Steps via MCP, Smarter Simulate, and Batch Send Fixes](https://molted.email/blog/whats-new-mcp-journey-steps-april-2026)
- [How Policy Engines Prevent Deliverability Disasters](https://molted.email/blog/how-policy-engines-prevent-deliverability-disasters)
- [What's New: Journeys in the Portal and Automatic Warmup Scheduling](https://molted.email/blog/whats-new-journeys-warmup-april-2026)
- [Suppression Lists and Bounce Handling for AI Senders](https://molted.email/blog/suppression-lists-and-bounce-handling-for-ai-senders)
- [One API Call: Send, Check, Deliver, Track](https://molted.email/blog/one-api-call-send-check-deliver-track)
- [Dunning and Billing Recovery with AI Agents](https://molted.email/blog/dunning-and-billing-recovery-with-ai-agents)
- [Multi-Provider Failover: Never Depend on One ESP](https://molted.email/blog/multi-provider-failover-never-depend-on-one-esp)
- [How AI Agents Send Email Today (and Why It's Broken)](https://molted.email/blog/how-ai-agents-send-email-today-and-why-it-s-broken)
- [Rate Limiting and Cooldowns: Protecting Sender Reputation](https://molted.email/blog/rate-limiting-and-cooldowns-protecting-sender-rep)
- [Molted is Now an MCP Server: Give Your Agent a Policy-Enforced Mailbox in One Line](https://molted.email/blog/molted-mcp-server)
- [Automated Churn Prevention with AI Agents](https://molted.email/blog/automated-churn-prevention-with-ai-agents)
- [How We Scan for Prompt Injection in Email Pipelines](https://molted.email/blog/how-we-scan-for-prompt-injection-in-email-pipelines)
- [What's New: Free Tier, Broadcast Send, and Full Mailbox Control](https://molted.email/blog/whats-new-april-2026)
- [Email Deliverability for AI Agents: A Technical Guide](https://molted.email/blog/email-deliverability-for-ai-agents-a-technical-guide)
- [Resend vs Postmark vs Molted: Which Email API for AI Agents?](https://molted.email/blog/resend-vs-postmark-vs-molted-which-email-api-for-ai-agents)
- [AI-Powered Outbound Prospecting That Doesn't Spam](https://molted.email/blog/ai-powered-outbound-prospecting-that-doesn-t-spam)
- [Autonomy Levels: From Human-Approved to Fully Autonomous](https://molted.email/blog/autonomy-levels-from-human-approved-to-fully-autonomous)
- [Building an Email Pipeline for AI Agents](https://molted.email/blog/building-an-email-pipeline-for-ai-agents)
- [The Policy Rules That Protect Your Sender Reputation](https://molted.email/blog/the-policy-rules-that-protect-your-sender-reputation)
- [How AI Agents Handle Inbound Replies](https://molted.email/blog/how-ai-agents-handle-inbound-replies)
- [AI Agent Onboarding Emails: The Complete Guide](https://molted.email/blog/ai-agent-onboarding-emails-the-complete-guide)
- [What Happens When an AI Agent Over-Sends](https://molted.email/blog/what-happens-when-an-ai-agent-over-sends)
- [How to Send Your First Policy-Checked Email](https://molted.email/blog/send-your-first-policy-checked-email)
- [What Is Agent-Native Email?](https://molted.email/blog/what-is-agent-native-email)
- [One-Click Domain Setup with Domain Connect](https://molted.email/blog/one-click-domain-setup-with-domain-connect)
- [Why AI Agents Need Email Guardrails](https://molted.email/blog/why-ai-agents-need-email-guardrails)

## Optional

- [Changelog](https://molted.email/changelog)
- [Support](https://molted.email/support)
- [Terms of Service](https://molted.email/terms)
- [Privacy Policy](https://molted.email/privacy)