The cognitive email layer for AI

Make your agents smarter by feeding them context, not just raw HTML.Agentry understands an inbox by analyzing content and mapping relationships the moment an email arrives.

No credit card required

Traditional APIs feed your AI messy HTML. Agentry feeds it context.

Webhook payload preview

Simulate incoming email

X-MS-Exchange-Organization-Network-Message-Id: 08d9a1f2-3c4b-56e7-8901-23456789abcd
X-Mailer: Microsoft Outlook 16.0
Thread-Topic: Integration outage + SLA
Thread-Index: AdZx9y0w8v7u6t5s==
MIME-Version: 1.0
Content-Type: multipart/alternative;
 boundary="----=_Part_982374_87654321.1711271640123"
From: "Sarah Jenkins" <sarah@enterprise.com>
To: "You" <you@company.com>
Cc: legal@enterprise.com, team@company.com
Subject: =?UTF-8?Q?URGENT:_Production_integration_down_=E2=80=94_SLA?=
Message-ID: <CA+xyz=abc987.enterprise.com>
Date: Mon, 24 Mar 2026 09:13:58 +0000
X-Priority: 1 (Highest)
X-MSMail-Priority: High
Importance: high

------=_Part_982374_87654321.1711271640123
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable

Hey =E2=80=94 the new update broke our integration. Are we still on for th=
e $15k SLA? Fix this ASAP.

------=_Part_982374_87654321.1711271640123
Content-Type: text/html; charset="utf-8"
Content-Transfer-Encoding: quoted-printable

<html xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-mic=
rosoft-com:office:office">
<head>
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dutf-8">
<!--[if gte mso 9]><xml><o:OfficeDocumentSettings><o:AllowPNG/></o:OfficeD=
ocumentSettings></xml><![endif]-->
<style type=3D"text/css">
<!--
.body { font-family: 'Segoe UI', Arial, sans-serif; color: #333333; }
.small { font-size: 11px; color: #888888; }
-->
</style>
</head>
<body class=3D"body">
<table role=3D"presentation" width=3D"100%" cellpadding=3D"0" cellspacing=
=3D"0" style=3D"border-collapse:collapse;mso-table-lspace:0pt;mso-table-rsp=
ace:0pt;">
<tr><td style=3D"padding:12px 0;">
<img src=3D"data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///ywAAAAAAQABAAACAUwAOw=3D=3D" width=3D"1" height=3D"1" alt=3D"" style=3D"display:block;border:0;outline:none;text-decoration:none;">
</td></tr>
<tr><td style=3D"font-size:15px;line-height:22px;">
Hey, the new update broke our integration.
</td></tr>
<tr><td class=3D"small" style=3D"padding-top:8px;">
Are we still on for the $15k SLA? Fix this ASAP.
</td></tr>
<tr><td style=3D"padding-top:24px;font-size:10px;color:#aaaaaa;">
This message may contain confidential information. If you are not the inten=
ded recipient, delete it. Tracking ref: ENT-2026-88421-MSG
</td></tr>
</table>
<!--[if mso]><table role=3D"presentation"><tr><td height=3D"0"><![endif]-->
<div style=3D"font-size:0;line-height:0;max-height:0;overflow:hidden;mso-hi=
de:all;">
<img src=3D"https://track.enterprise.com/pixel.gif?m=3D88421&u=3D9f3e2d1c" width=3D"1" height=3D"1" alt=3D"">
</div>
</body>
</html>

------=_Part_982374_87654321.1711271640123--

Agent status

[!] Consuming 12,400+ tokens…

[!] Parsing quoted-printable, MSO conditionals, and nested tables…

[!] Context window pressure — headers + HTML before the actual ask.

One cognitive layer for bots and the people they work with

Give your AI its own inbox, or connect it to a human's mail — same pipeline, so your agents stay aligned with real conversations without rebuilding everything twice.

Autonomous agents

Give your agents their own inboxes at @yourcompany.agentry.to to send and receive email autonomously.

Agentryunderstanding layer

Email is normalized, summarized, and tied to memory graphs your agent can actually use.

Human mailboxes

Connect your own mailboxes — Gmail via Google sign-in or any standard IMAP server. Now every time a message arrives, your agent gets structured context alongside the cleaned content.

OAuth · IMAP · secure credentials

Built for agents that need to be right, not just loud

Engineering choices that make email a first-class input for helpful, grounded assistants.

Smart webhooks, full inbox behind them

Stop paying models to wade through newsletters and boilerplate. Declare what matters to each agent — we shape webhook payloads so noise stays light and signal stays high.

One brain for bots and humans

Autonomous agents and human copilots share the same layer, so context does not fracture. A support bot can reason about how a founder usually talks to a customer; handoffs from person to agent stay coherent because the story lives in one place.

Draft-first sending

Keep humans in the loop when stakes are high. Push the agent's proposed reply into the mailbox as a draft instead of sending blindly — one click to approve, edit, or discard before anything goes out.

Private by default. Built for compliance-minded teams.

AI on real inboxes touches regulation and trust. We focus on data minimization in what we send to your agents, clear deletion paths, and fewer surprises for your security review.

Full history when you need it

We retain message content so you can retrieve any thread or attachment over the API. Webhooks can carry rich summaries while the originals stay available for deeper lookups.

One-call "forget me"

A single delete on a contact cascades through stored mail, embeddings, and scores so people can exercise their rights without you wiring custom purge jobs.

Your keys, your models

Our provider APIs do not train on your content. But you can bring your own keys when you want full control over where inference runs (coming soon).

Stop building email parsers. Start building intelligence.

First-party SDK is on the way. Until then, use the AgentMail-compatible client pointed at Agentry — same patterns, our API.

Coming soon

npm install @agentry/sdk

Predictable pricing — pay for the compute you use, not the spam you don't.

See plans