Product
A dedicated address, a thread history, and a console a human can read — without handing an agent your own mailbox.
An AI agent that emails from your personal mailbox inherits your identity, your contacts and your blast radius. A dedicated inbox gives it an address of its own: mail arrives there, the agent reads it through the API or an MCP tool, and every reply is attributable to the agent rather than to you.
Every project is either live or test, and a message never crosses between them by changing a setting.
An address on agents.emailforagents.ai that sends and receives real
internet mail. Early access includes one complimentary live inbox and 100 email units a
month per verified owner identity, claimed by that identity's first eligible workspace.
Creating a second workspace does not renew it.
Non-routable .mail.invalid addresses behind a fake mail transport, so they
cannot reach the internet at all. Simulate an inbound message, let the agent draft a
reply, approve it, and read the events — with no recipient anywhere. Up to 3 active test
inboxes, the last 100 messages, 10 MB, and message content expires after 7 days.
The console is not an afterthought on top of an API. It is where you find out what your agent has been saying.
Threads, not rows. Replies land back on the same conversation. Reply to one message, or to everyone already on the thread.
Honest statuses. Queued, accepted by the provider, delivery updated, failed — kept distinct, because provider acceptance is not proof that a person received anything.
Search that says what it covers. Stored subjects, senders and previews, including conversations past the loaded page. Full message-body and attachment-content search are not included.
Files with a gate. Attachments become available only after upload and safety checks pass. Pending or quarantined files stay unavailable rather than silently downloading.
These are enforced in code, not aspirations. An agent that exceeds one gets an error, not a surprise.
Up to 25 unique recipients, a 998-character subject, and 512,000 characters of body text.
Up to 10 attachments, 10 MB each and 15 MB in total. Inbound MIME is accepted up to 25 MB.
3 active test inboxes, the most recent 100 messages, and 10 MB of retained content.
Message content expires 7 days after it arrives. Nothing is ever sent to the internet.
One complimentary live inbox and 100 email units a month per verified owner identity. One unit is one incoming message, or one recipient a provider accepts.
When the allowance runs out, sending stops. There is no overage and no surprise invoice.
Platform addresses work from the first minute. A domain of your own is a paid-plan feature, and Stripe checkout is disabled, so it cannot be bought yet.
When it opens, the DNS wizard separates the three jobs — proving ownership, sending, and receiving — and preserves your existing mail routing unless you explicitly choose to migrate it. That distinction matters: pointing MX records at a new provider by accident is how a company loses a day of mail. Read the domain setup guide.
Agent email thread ID vs message ID vs the Message-ID header: which to store, which a counterparty can forge, and what reply_to_message_id accepts.
DMARC for AI agent senders: why a fresh agent subdomain should start at p=reject, when that flips, and what sp, np and alignment actually cover.
Why does my AI agent email go to spam? The three agent-specific causes, what the platform reports back, and the part no sender can see.
The recipient rules an agent should follow on reply-all: who goes in To, what to keep in Cc, why Bcc never carries forward, and what each extra name costs.
How many tokens does an email thread use? We built a 10-message thread and counted: 4% was new text. Here is the retrieval pattern that fixes it.
Strip quoted reply text before an email agent reads it: client quote markers, a parsing order, and the replies where the quote is the message.
A timeout pattern for an agent waiting on an email reply: three states, the events that move them, in-thread nudges, and defaults by context.
How to let my agent read its email: list threads, page messages, fetch a body, and the ordering rules that decide if it sees a reply before it acts.
Sending returns HTTP 402, not 429. What keeps working, how a unit is counted, which reservations stay held, and the upstream quota you do not control.
Run agent mail on agents.yourcompany.com: inspect existing MX, publish ownership, SPF, DKIM and receiving records, and leave employee routing untouched.
A working pattern for support triage: one dedicated address, read and draft permission, approval on anything money-related, and events that prove it worked.
Giving an agent its own inbox versus wiring it into an existing Gmail or Microsoft 365 mailbox: identity, blast radius, OAuth scope and revocation.
Agents that create accounts need a real mailbox for verification email. How to provision one, read the code safely, and where this becomes unacceptable use.
Receive invoices at a dedicated address, scan attachments before the agent acts, draft the confirmation, and require approval before anything is sent back.