COMING SOON · FROM THE KIRIM.EMAIL TEAM
One API call, and your AI Agent has an email address of its own. Incoming mail gets checked before your AI Agent reads it. Replies that touch money, contracts, or sensitive data wait for your approval. Every send is logged: which AI Agent, when, and on whose behalf.
One agent. Its own identity.
cs-agent@contoh.email.ai.id
Every step, on record.LOG / 03
01020369% of companies still run AI Agents on shared credentials. Only 32% give every AI Agent an identity of its own.
VentureBeat survey, June 2026, of 107 companies
Runs on KIRIM.EMAIL's delivery infrastructure, used by 35,000+ businesses in 20 countries.
THE PROBLEM
Today, an AI Agent that needs email takes one of three shortcuts. All three work fine, until one day they don't.
AI Agents use an employee's Gmail or Workspace account. Two-factor checks and sending limits break the automation.
Suspend the account, and the agent stops too. Logs record its emails under the employee's name, not the agent's identity.
Hidden instructions inside an email can tell an agent to leak data or send a reply without your approval.
The agent reads everything, including text you cannot see. The threat stays out of your sight.
Auditors ask who sent an email and where its data is stored. Without an identity and audit trail, teams struggle to answer.
Today's agent inbox services store data overseas, without local contracts or local support.
This already happened: in 2025, researchers showed a single email could make Microsoft 365 Copilot leak internal data with no click from the user (EchoLeak, CVE-2025-32711).
So AI Agents get delayed, or they run with risk that only shows up after the fact.
HOW IT WILL WORK
Here's how it will work when we launch.
One API call, or one command in your terminal. Your AI Agent has its own email right away.
Incoming mail goes through Agent Firewall first, then reaches your AI Agent as JSON over a webhook.
Safe replies go straight out. Risky ones queue for approval. You just tap Approve or Edit from WhatsApp or Slack.
# Draft example, not final
curl -X POST https://api.email.ai.id/v1/inboxes \
-H "Authorization: Bearer ***" \
-d '{"name": "cs-agent"}'
# Response
{
"inbox_id": "inb_7Hq2",
"address": "cs-agent@example.email.ai.id",
"data_region": "indonesia"
}The example above is still a draft. We'll publish the final API spec at launch.
WHAT WE ARE BUILDING
Coming soonIncoming mail gets checked before your AI Agent reads it. Hidden text is stripped out, prompt injection patterns are flagged, and you can limit who is allowed to email each inbox.
INCOMING EMAIL CHECKSConcept illustration
email → screening → agent
Replies that touch money, contracts, or sensitive data are held as drafts until you approve or edit them from WhatsApp or Slack.
Sending paused
WhatsAppSlack
human-in-the-loopEvery send is logged: which AI Agent, on whose behalf, when. Data and logs stay in Indonesia, with local contracts, invoices, and support.
One call to create an inbox. Threads stay intact, replies are split out from quoted text, and attachments turn into structured data.
POST/v1/inboxes
{ "name": "cs-agent" }You don't have to build your own inbound email pipeline, and you don't have to buy a Workspace license for every AI Agent.
WHO IT IS FOR
Banks, fintechs, state-owned enterprises, government agencies, and corporations that must account for every AI Agent email.
Teams building AI Agents who don't want to spend weeks building their own inbound email pipeline.
WHY KIRIM.EMAIL
email.ai.id runs on the same delivery path 35,000+ businesses use today. We have spent years on the hardest part of email: getting messages into the inbox instead of the spam folder. Now we are bringing that experience to AI Agents.
FAQ
We're still building it. Access opens in waves, starting with the people on this waitlist. We'll email the launch date to everyone on the list.
email.ai.id is a new product from KIRIM.EMAIL (PT Kirim Email Indonesia). It runs on the same email infrastructure KIRIM.EMAIL customers already use today.
Pricing isn't final yet. We're planning a free tier so you can try it out, plus pay-as-you-go. People on the waitlist will hear about pricing first.
Design partners get the earliest access, direct sessions with the product team, and a say in what we build first. In return, we ask for regular feedback during the pilot.
Services like AgentMail are built for global developers, with data stored outside Indonesia. email.ai.id is built for teams that have to account for every AI Agent send: incoming mail is screened, risky replies are held, every send is logged, and the data stays in Indonesia with local contracts and invoicing.
Not for every kind of data. GR 33/2026, in force from 16 January 2027, sets the conditions for transferring personal data abroad. With your data stored in Indonesia, your AI Agent email involves no cross-border transfer, which makes audits and procurement simpler. For obligations specific to your sector, check with your legal team.
No, and we won't promise that. No filter catches every attack. That's why there are two layers: the firewall flags and cleans suspicious email, then Approval Gate holds risky replies until you approve them.
Sign up now. Teams already running AI Agents hear from us first.
Access opens in waves. Teams already running AI Agents hear from us first.
No credit card. We only email you about email.ai.id.
By signing up, you agree that we can contact you about email.ai.id. See our Privacy Policy.