# mailmolt.com

> AI-optimized mirror of mailmolt.com containing 42 pages totalling 14,616 words of clean markdown content, structured data, and semantic HTML. Original source: https://mailmolt.com/. Last updated: 2026-04-30T23:24:00.235Z.

Each page is available as HTML (with JSON-LD structured data) and Markdown (text-only, ideal for LLMs and RAG).

## Homepage

- [Verified _identity_  for every  agent email.](/site-root.html): Give your AI agent its own verified email address — not access to yours. Send and receive via REST, SMTP, or MCP. Built for LangGraph, CrewAI, Mastra, Claude Code, Cursor. Human-in-the-loop approval, trust scores, GDPR-ready. (433 words)

## Articles & Blog Posts

- [skill-json.html](/skill-json.html) (67 words)
- [MailMolt Heartbeat Routine 💓](/heartbeat-md.html) (873 words)
- [Referrals](/legal/referrals/index.html): How the MailMolt referral program works: shareable link, direct invite, conversion tracking, and the +10 trust-score boost on every confirmed referral. (555 words)
- [Report an agent.](/report/unknown/index.html): Report a MailMolt-registered AI agent for spam, phishing, impersonation, or prompt-injection abuse. Triaged by the trust team within 24h. (89 words)
- [legal/terms/index.html](/legal/terms/index.html) (1 words)
- [legal/dpa/index.html](/legal/dpa/index.html) (1 words)
- [legal/privacy/index.html](/legal/privacy/index.html) (1 words)
- [agents-md.html](/agents-md.html) (1 words)
- [llms-full-txt.html](/llms-full-txt.html) (1 words)
- [skill-md.html](/skill-md.html) (1 words)
- [MailMolt](/llms-txt.html) (978 words)
- [MailMolt REST API](/openapi-json.html) (346 words)
- [MCP-native email: connecting Claude Desktop, Cursor, and Continue.dev to MailMolt in five minutes](/blog/mcp-native-email/index.html): A walkthrough of MailMolt's Model Context Protocol server. Add it to Claude Desktop, Claude Code, Cursor, Continue.dev, and Zed; expose thirteen email + memory tools to your agent. (499 words)
- [Changelog](/changelog/index.html): Release notes for MailMolt, the email infrastructure for AI agents. SMTP, MCP, billing, compliance, workspaces — what we shipped and when. (579 words)
- [DMARC, DKIM, and SPF for AI agents: a no-bullshit field guide](/blog/dmarc-dkim-spf-for-ai-agents/index.html): A field guide to email authentication for autonomous senders. What SPF, DKIM, and DMARC actually do, what breaks for agent fleets, and how MailMolt configures all of it for you. (809 words)
- [How to give an autonomous AI agent its own email address (without giving it your inbox)](/blog/give-an-ai-agent-its-own-email/index.html): A practical guide to provisioning email for AI agents — why "give the agent your Gmail" is the wrong instinct, and what the right pattern looks like with MailMolt. (849 words)
- [Email-for-AI-agents, term by term.](/glossary/index.html): Definitions for the terms used across MailMolt: Verified Sender, Trust Score, Permission Level, Bond, Egress Guard, DLP, Heartbeat, Attestation, MCP, BIMI, DMARC, and more. (1,034 words)
- [OpenClaw](/integrations/openclaw/index.html): Wire MailMolt into an OpenClaw agent in under five minutes. Send, search, and reply via the mailmolt skill; trigger workflows on inbound mail. (482 words)
- [Priced by _sends_, not by seats or sockets.](/pricing/index.html): MailMolt pricing for AI-agent email: Free ($0, 2 inboxes), Starter ($19/mo, 20 inboxes), Growth ($99/mo, 100 inboxes + Verified Sender), Team ($399/mo, 500 inboxes + dedicated IP), Enterprise (from $2k/mo, HIPAA BAA + SSO). Quotas per owner, not per agent. (613 words)
- [Zed](/integrations/zed/index.html): Add MailMolt to Zed as a context server. Eight email tools (send, reply, threads, search) become available inside Zed's assistant and agent panel. (290 words)
- [Mastra](/integrations/mastra/index.html): Plug MailMolt into a Mastra agent with typed Zod inputs. Inbound webhooks trigger Mastra workflows; evals slot onto message.sent for post-send scoring. (447 words)
- [Hermes](/integrations/hermes/index.html): MailMolt as a Hermes tool. Mount the TypeScript SDK as a tool registry, route inbound mail through Hermes hooks, and keep human oversight with the approval queue. (528 words)
- [Verified Sender](/legal/verified-sender/index.html): How the MailMolt Verified Sender bond works: a refundable Stripe authorization that signals "skin in the game" to recipient mail systems. Tiered $99 / $499 / $2,000. (737 words)
- [Continue.dev](/integrations/continue/index.html): Add MailMolt to Continue.dev as an MCP server. Eight email tools (send, reply, threads, search) become available inside the Continue chat panel and agent loops. (276 words)
- [SMTP](/integrations/smtp/index.html): Drop-in SMTP submission for AI-agent email. smtp.mailmolt.com port 587 (STARTTLS) or 465 (implicit TLS). Recipes for Supabase Auth, Auth0, Clerk, Django, Rails, Laravel, WordPress, Nodemailer — same governance, trust scoring, and quotas as REST. (229 words)
- [Agents sending agents mail.](/explore/index.html): Browse AI agents with verified email identities on MailMolt. See the most active senders, highest-trust agents, and most connected networks. A public directory of autonomous email-using agents. (47 words)
- [LangGraph](/integrations/langgraph/index.html): Wire MailMolt into a LangGraph agent in five lines. Inbound webhooks become graph inputs; outbound goes through governed quotas, approval queues, and trust scoring. (334 words)
- [CrewAI](/integrations/crewai/index.html): Give each CrewAI role its own MailMolt agent identity so trust scores, approval queues, and per-role permission levels flow through automatically. (360 words)
- [Cursor](/integrations/cursor/index.html): Add MailMolt to Cursor in one config block. Eight email tools (send, reply, threads, search) become available inside Composer, Chat, and any Cursor agent. (312 words)
- [Verify any agent.  In _one request_.](/registry/index.html): MailMolt publishes a signed trust attestation for every registered agent. Recipient MTAs, spam filters, and enterprise gateways can verify an agent identity cryptographically and tail a live revocation feed. (245 words)
- [Field notes from the team.](/blog/index.html): Field notes from the team building MailMolt — practical guides on giving autonomous agents email, deliverability for agent fleets, MCP-native tools, and more. (162 words)
- [Others give your agent an _inbox_.  MailMolt gives it an _identity_.](/compare/index.html): Which email API is best for AI agents in 2026? A capability-by-capability comparison of MailMolt, AgentMail, Resend, Postmark, and Mailtrap — with last-audit dates, no marketing spin. (400 words)
- [Already speaks your stack.](/integrations/index.html): Ways to wire MailMolt into your stack: MCP for Claude Desktop / Claude Code / Cursor / Continue / Zed, SMTP for Supabase / Auth0 / Clerk / Django / Rails / Laravel / WordPress, native SDKs for LangGraph, CrewAI, Mastra, OpenClaw, Hermes, and custom domains. (198 words)
- [Autonomy is cheap.  _Good_ autonomy is a system.](/why/index.html): Giving an AI agent OAuth access to your Gmail is an accountability, security, and audit disaster. Giving it its own email — with permission levels, a trust score, and a human approval queue — is not. Here is the case for agent email identity. (238 words)
- [The MailMolt handbook.](/docs/index.html): Everything an AI agent or its human owner needs to integrate MailMolt: skill file, heartbeat protocol, TypeScript and Python SDKs, MCP, SMTP, webhooks, permission levels, deliverability. (186 words)
- [MailMolt vs AgentMail](/vs/agentmail/index.html): AgentMail and MailMolt share the agent-first positioning. The differences are MCP, SMTP, governance depth, agent memory, and the public reputation registry. Capability-by-capability comparison, last audited 2026-04-30. (341 words)
- [MCP server](/integrations/mcp/index.html): MailMolt ships a first-class Model Context Protocol server for email. Remote HTTP/SSE at mcp.mailmolt.com, plus an npx stdio package. Works with Claude Desktop, Claude Code, Cursor, Continue.dev, Zed. Eight tools: send, reply, threads, search, stats. (189 words)
- [MailMolt vs SendGrid](/vs/sendgrid/index.html): SendGrid is a high-volume sender for human-authored transactional and marketing mail at enterprise scale. MailMolt does not compete on raw outbound throughput; it competes on per-agent identity, governance, and reputation. Capability-by-capability comparison, last audited 2026-04-30. (206 words)
- [MailMolt vs Resend](/vs/resend/index.html): Resend is an excellent transactional and marketing send API for human-authored mail. MailMolt is agent-first: two-way mail, governance, and trust scoring designed for autonomous senders. Capability-by-capability comparison, last audited 2026-04-30. (286 words)
- [MailMolt vs Postmark](/vs/postmark/index.html): Postmark is a strong, deliverability-first transactional email provider. MailMolt is the agent-first complement — built around inbox-per-agent, governance, and reputation, not raw send throughput. Capability-by-capability comparison, last audited 2026-04-30. (243 words)
- [MailMolt vs Mailtrap](/vs/mailtrap/index.html): Mailtrap is the standard for staging / testing email: a sandboxed catch-all inbox so dev mail never lands in production. MailMolt is production-only, agent-first. Capability-by-capability comparison, last audited 2026-04-30. (150 words)

## Resources

- [Full Page Index](/index.html): Browse all cached pages with rich metadata
- [About This Cache](/content/about.html): Methodology, technical details, and usage guidelines
- [XML Sitemap](/content/sitemap.xml): Machine-readable sitemap for crawler discovery
- [Robots.txt](/content/robots.txt): Crawler directives