Email API built for AI agents
Let your agents read, send, and react autonomously, without human involvement
AgentSkill
MCP
Bash
Copy
# Register
npx --package=@atomicmail/agent-skill atomicmail register --username "myagent"

# Send a JMAP request inline
npx --package=@atomicmail/agent-skill atomicmail jmap_request \
  --ops '[["Mailbox/get", {"accountId": "$ACCOUNT_ID"}, "m0"]]'

# Send a JMAP request from a preset file
npx --package=@atomicmail/agent-skill atomicmail jmap_request --ops-file send_mail.json

# Ask for help
npx --package=@atomicmail/agent-skill atomicmail help
JSON
Copy
{
  "mcpServers": {
    "atomicmail": {
      "command": "npx",
      "args": ["-y", "@atomicmail/mcp"]
    }
  }
}

Free accounts

We are on open alpha. The service is 100% free.

Custom domains

Use @atomicmail.ai
or your custom domain

Any agent is supported

3 things no other mail offers

/01

Let agents create their own email

No identity verification, no credit cards, no domains to configure. Your agent requests a real inbox via API and gets to work instantly.

/02

AI-native JSON API

Powered by standard JMAP and REST. Cleanly structured so LLMs can generate correct API calls natively, without complex vendor SDKs.

/03

Agents that learn from mistakes

Built for LLM error recovery. Every API error returns a plain-language hint, allowing your agent to debug and fix its own requests autonomously.

Standards built for AI agents

POW & Reputation

Each connection is verified with a Proof-of-Work challenge, cheap for one agent, prohibitive for a million. Each agent also earns a reputation score that grows with every successful, non-flagged interaction. Good agents move fast. Bad ones hit a wall.

MCP, AgentSkill or JMAP API

Via MCP for chat-based agents (Claude Desktop, Cursor, etc.), via AgentSkill for shell-capable agents, or via direct API if you want to manage auth yourself, run outside Node/Deno/Bun, or integrate without MCP/skill wrappers.

Docs

Use cases

Newsletter intelligence

Subscribe an agent inbox to dozens of industry newsletters. The agent reads everything, extracts signals, and surfaces only the items that match a user-defined interest profile — once a day, as a digest. Zero noise in the user's personal inbox.

Autonomous job applications

An agent finds job listings, drafts tailored applications, sends them from its own inbox, monitors replies, and escalates to the human only when a recruiter responds with a real opportunity.

Vendor invoice processing

An agent inbox receives supplier invoices by email. The agent parses each one, matches it against a PO database, and routes exceptions to a human approver — all without touching the company's main email domain.

Async user research interviews

An agent conducts email-based user research: sends structured questions, reads freeform responses, follows up based on what it learns, and summarizes findings. Respondents reply on their own schedule — no scheduling, Calendly, or Zoom.

SaaS account provisioning

Subscribe an agent inbox to dozens of industry newsletters. The agent reads everything, extracts signals, and surfaces only the items that match a user-defined interest profile — once a day, as a digest. Zero noise in the user's personal inbox.

Competitive monitoring

An agent subscribes to competitor product update emails, release notes, and press releases. It tracks changes over time, maintains a diff, and alerts the user when a competitor ships something significant.

Customer support at the edge

A support agent owns 'support@' and handles inbound tickets end-to-end: reads the issue, queries internal knowledge bases, and replies with a complete answer. Humans receive only the escalations the agent cannot resolve.

Price and availability tracking

An agent finds job listings, drafts tailored applications, sends them from its own inbox, monitors replies, and escalates to the human only when a recruiter responds with a real opportunity.

Agent-to-human escalation layer

Any agentic pipeline can send its owner a plain email when it encounters an edge case requiring judgment. The human replies in natural language; the agent parses the reply and resumes. This turns email into a universal human-in-the-loop interface.

Regulatory & compliance notifications

An agent monitors regulatory sources, receives filings and updates via mailing lists, extracts the relevant changes, and emails a structured briefing to the compliance team. Cheaper than a dedicated compliance SaaS.

AI pen pal / communication layer

Two people or organizations communicate through AI intermediaries — each party has an agent that reads messages, drafts responses for human approval (or sends autonomously with a trust level set by the user), maintains conversation.

Comparison with alternatives
Capabilities
Atomic Mail

Agent-native

AgentMail

Agent-native

Resend

Dev-friendly

SendGrid / Mailgun

Traditional ESP

Designed for agents

PRIMARY

PRIMARY

SECONDARY

HUMANS FIRST

Agent registers autonomously

POW ONLY

VERIFICATION

ACCOUNT + DOMAIN

MANUAL CREATION

Full inbox API (receive emails)

FULL JMAP INBOX

REST INBOX API

SEND ONLY

SEND ONLY

LLMs know the API natively

JMAP IS IN
TRAINING DATA

MUST LEARN CUSTOM SDK

MUST LEARN CUSTOM SDK

MUST LEARN CUSTOM SDK

Self-documenting errors & hints

hint, docs_url, _next

NO

NO

NO

MCP server

DEDICATED

DEDICATED

NO

NO

AgentSkill

SUPPORTED

NO

NO

NO

No domain required

YES

YES

NO

NO

Standards-based (no vendor lock-in)

JMAP IS AN IETF
STANDARD

PROPRIETARY

PROPRIETARY

PROPRIETARY

API protocol

JMAP (RFC 8620/8621)

PROPRIETARY REST

PROPRIETARY REST

PROPRIETARY REST /
SMTP

/01

Designed for agents

Atomic Mail

PRIMARY

AgentMail

PRIMARY

Resend

SECONDARY

SendGrid, Mailgun

HUMANS FIRST

/02

Agent registers autonomously

Atomic Mail

POW ONLY

Agent Mail

VERIFICATION

Resend

ACCOUNT+DOMAIN

SendGrid, Mailgun

MANUAL CREATION

/03

Full inbox API (receive emails)

Atomic Mail

FULL JMAP INBOX

AgentMail

REST INBOX API

Resend

SEND ONLY

SendGrid, Mailgun

SEND ONLY

/04

LLMs know the API natively

Atomic Mail

JMAP IS IN
TRAINING DATA

AgentMail

LEARN CUSTOM SDK

Resend

LEARN CUSTOM SDK

SendGrid, Mailgun

LEARN CUSTOM SDK

/05

Self-documenting errors & hints

Atomic Mail

hint

docs_url

_next

AgentMail

NO

Resend

NO

SendGrid, Mailgun

NO

/06

MCP server

Atomic Mail

DEDICATED

AgentMail

DEDICATED

Resend

NO

SendGrid, Mailgun

NO

/07

AgentSkill

Atomic Mail

SUPPORTED

Agent Mail

NO

Resend

NO

SendGrid, Mailgun

NO

/08

No domain required

Atomic Mail

YES

AgentMail

YES

Resend

NO

SendGrid, Mailgun

NO

/09

Standards-based (no vendor lock-in)

Atomic Mail

JMAP IS AN
IETF STANDARD

AgentMail

PROPRIETARY

Resend

PROPRIETARY

SendGrid, Mailgun

PROPRIETARY

/10

API protocol

Atomic Mail

JMAP (RFC 8620/8621)

AgentMail

PROPRIETARY REST

Resend

PROPRIETARY REST

SendGrid, Mailgun

PROPRIETARY REST

FAQ

Can an agent register an inbox without any human involvement?
/+/
Yes. The only requirement is the ability to solve an `scrypt` PoW challenge, which currently takes roughly 30 seconds on a standard inference server. No email confirmation, no domain, no credit card, no CAPTCHA.
Does Atomic Mail require me to own a domain?
/+/
No. Every inbox is on `@atomicmail.ai`. Custom domains are not supported in the alpha.
Can I use Atomic Mail from any language or runtime?
/+/
Yes. The JMAP API is JSON over HTTPS. Any HTTP client works. The MCP and AgentSkill packages add convenience for Node, Deno, and Bun environments, but they are not required.
Is JMAP hard to learn?
/+/
Most LLMs already know JMAP well enough to write correct method calls without documentation. If your agent is uncertain, `help` returns a full cheatsheet and worked examples.
What if I'm using a weaker model that doesn't know JMAP?
/+/
Semantic tools (`send_mail`, `get_emails`, `reply`, `search`, `mark_read`) are shipping soon as higher-level wrappers over `jmap_request`. Until then, use the preset system — shipped presets cover the common operations without requiring the model to generate JMAP itself.
What happens when the alpha ends?
/+/
Accounts created during the alpha are migrated to the free tier of the paid product. No data loss; no forced re-registration.
Can the MCP server and the AgentSkill be used simultaneously?
/+/
Yes. They share state. Both will see up-to-date credentials and tokens regardless of which one last refreshed them.
What if my agent gets stuck?
/+/
Call `help`. If an error response didn't include a useful `hint`, that's a bug — please report it.