Description
MachineMail is an email API tailored for AI agents, enabling them to send, receive, and manage email communications securely. It provides every agent with its own email address, eliminating inbox fees and offering a free tier that includes 3,000 messages per month. The core of MachineMail's offering is its advanced trust layer, which analyzes every inbound message before it reaches the agent. This layer combines SPF, DKIM, and DMARC alignment, sender history, allow/block lists, and checks for hidden instructions, links, and attachments, assigning a trust level to each message. This ensures agents interact only with agent-safe text, preventing prompt injection, phishing, and other manipulative content.
Read More
The platform is built for agents, supporting self-registration through an `auth.md` flow where an agent can obtain a sandbox inbox and an access token, which a human can later claim. It offers flexible inbox creation, either through a single API call or via catch-all domains that automatically create an inbox upon the first delivery. MachineMail is AgentMail-compatible, meaning existing clients can switch by simply changing the base URL and API key, as its `/v0` API follows AgentMail's paths, field names, and enums.
Delivery options include webhooks for pushing mail to an agent's stack and a hosted MCP (Machine Communication Protocol) server, allowing agents to pull mail. Webhooks are signed, retried for five days, and replayable from the dashboard, providing reliable event delivery for `inbox.created`, `message.received`, and `message.bounced` events. The MCP server supports `wait_for_message` calls that hold open until mail arrives.
MachineMail offers various pricing plans beyond the free tier, including Pro ($9/month), Team ($49/month), and Scale ($199/month), which scale message allowances, storage, custom domains, and support options. All plans include unlimited inboxes and a trust report on every message. Overage charges apply for messages, storage, and custom domains beyond plan limits, ensuring continuous service for agents in production. The service emphasizes security and reliability, providing a dedicated email infrastructure where AI agents can operate with confidence.
- CategoryAI Agent Builders
- Pricing ModelFreemium
- DeploymentCloud
- PlatformsWeb
MachineMail's Core Features
Free Email Inboxes for AI Agents
Advanced Trust Layer on Every Inbound Message
AgentMail-Compatible API
Agent Self-Registration with Sandbox Inboxes
Catch-All Domains for Dynamic Inbox Creation
Webhooks for Real-Time Event Delivery
Hosted MCP Server for Message Pulling
Custom Domain Support
Scalable Pricing Plans with Overage Options
Detailed Trust Reports with Five Trust Levels
Agent-Safe Text Processing (Removes Hidden Instructions)
API Key Management and Scoped Permissions
Message Delivery Replay from Dashboard
Export Inbox Data as mbox or JSONL
CLI Tool for API Scripting
How to use MachineMail?
Create Account: Sign up for a free MachineMail account to get started.
Get API Key: Generate API keys from your dashboard under 'API keys'.
Register Agent: Paste the agent prompt into your MCP client (e.g., Claude Code, Cursor) to allow the agent to self-register and get a sandbox inbox.
Claim Agent: Enter the six-digit code provided by your agent at `/claim` to bind it to your organization.
Create Inbox: Use one API call to create an inbox for your agent, or configure a catch-all domain.
Send & Receive: Your agent can send and receive emails, with inbound messages processed by the trust layer and delivered as agent-safe text.
Monitor Events: Utilize webhooks or MCP to receive real-time events and messages for your agent's inbox.
MachineMail's Use Cases
- Automated Customer Support
- Agent-Based Sign-up Flows
- Secure Internal Communications
- Data Collection & Processing
- Automated Outreach & Replies
- Ephemeral Task-Specific Inboxes
- Prompt Injection Prevention
- Developer Tool Integration
