i love what workos is doing and i also spoke at their event already. we think this is complementary - we're not setting out to create a new standard, we're giving better access to what already exists.

we extensively rate limit and prevent outbound sending until an agent has been verified by a human :) more detailed in OG post

i appreciate your feedback and thanks for using agentmail!

thanks for the feedback! it is top of mind for us, and we've done our research to prevent this as much as possible. it is constantly ongoing and we never settle in our measures to protect against this

This was likely a free tier user. We do this intentionally and don't allow free users to send from custom domains, so you can have a easier time identifying LLM emails. In this case, it seemed like it worked :)

And for paid users the receivers don’t need to have an easier time identifying the LLM email? What kind of reasoning is that?

To be the advocate of the devil here:

A lot of people believe that spam issue would be largely solved if each email costed 0.001$

Appreciate the concern Mike, and I actually read your email complaining, which helped us ship this next feature. We have a "sent via AgentMail" footer being added soon to outbound emails to identify emails coming from LLM's.

We also are working on adding more robust checks and LLM-based filtering to prevent messages which contain spam or outbound-like copy.

Re; AgentMail next to Claude, we're working on stateful inboxes which help agents actually recall and understand what they're sending and to who. The goal is to provide the rails for intelligent actors rather than slop.

So, a footer to make sure they've already engaged with the content in good faith before seeing the spam warning, and which doesn't actually explain that the content is AI generated?

Just go post on black hat forums. Plenty of people want this, it's a spam service. You don't need to be here.

Re "sent via AgentMail" - that's good to hear, but I hope it's not the entire planned text, as "AgentMail" will mean nothing to most people that receive an email from your service. It wont indicate that the email was composed by an AI rather than a person, which is the information that needs to get across.

What if a people send email via te agent and using agent mail? writing by the agent but aproval by human.

There are OS-native options everywhere to spawn an email client window that's filled out and ready to send from your address, so that couldn't possibly be a differentiator for them

yep we're going to have a footer linked to our website, which should allow people to see that we are an email service for agents. thanks!

Can you not just make the text more descriptive? „Sent by a generative AI model“ or something? Nobody is going to click a link in a spam email.

He's obviously trying to avoid having to do this, so linking out to the website is the "hey look we're doing the right thing" when really they aren't.

Chances are more people would identify the service as something to block or report for spam if the text were more descriptive, so he's counting on people not clicking the link in the footer but at least he can claim it's there, even if it's ineffectual.

This service will be marked as spam anyway, it's only a matter of time.

They’re literally using an LLM to write the email.They could make the disclaimer text detailed and descriptive per email sent. So the use case they’re citing here that “well what if it’s just an email that was forwarded by the bot” doesn’t apply because you could add a different disclaimer message at the bottom of that email vs the spam example above.

Lead with the AI being sent by AI/Agent using the service.

Ban any sender using your domain that removes, obscures, hides, or alters this first line.

This response is a failure to understand the issue.

It's very hard to get someone to understand something if their salary depends on them not understanding it.

Oh he understands it, he just DGAF.

"sent via AgentMail" - removable by a higher plan later on

What is the point of automating the signup process?

It's less work to signup a second email address for agent use than to signup with you, then signup a second email address.

After all, it's not like each agent needs their own email.

Make sure it adds a header too

What kind of UserAgent is being used? This could be easily used in good faith by the sending party, then any spam blockers can remove everything from said UserAgent. If they then change their UserAgent to something generic, you know who's acting in bad faith.

agreed from a fundamental level. but i think being an intelligent and aware as an autonomous entity requires capabilities beyond sending. agents will need to have contextual awareness of the messages they send and receive

IMAP?

AgentMail provides the IMAP+SMTP server as managed service

nice, we're working on agentmail so i think we have similar ideas. its a big problem!

That discomfort of letting an agent touch your personal inbox was actually the core insight for me. You lose visibility into what it's doing, and mixing agent actions with human mail is just a recipe for confusion. An agent should have its own identity, its own address, its own inbox. Would love to see where you're taking agentmail — are you focused more on the agent-as-sender side, or agent-as-receiver (verification codes, inbound parsing, etc.)?

how are you implementing agentmail? would love to know more

I'm building a platform of AI agents, and each agent can have its own email address. AgentMail handles that. You create an inbox via their REST API and they POST to your webhook when mail arrives.

On the agent side, it just gets tools: send email, reply to email, list inbox, read message. Those tools call the AgentMail API. So the fake implements the same interface.. same send/reply/list/read methods, but recording calls instead of making HTTP requests. You can pre-populate inboxes with test messages, simulate "username taken" errors, etc.

AgentMail is actually one of the simpler fakes because there's no internal state to maintain. Sending a message doesn't affect what you'd read back from your own inbox. Some of the other fakes (like the database or file storage) need to actually simulate state in memory so writes are visible to subsequent reads. This one is closer to a stub.

thank you, this means a lot!

We know the founder of this pretty well actually. It's not really an alternative. They are an MCP that is building an email-like layer for coding agents to talk to each other. We are an actual email inbox provisioned via API for agents to email agents or humans.