エージェント向けに設計する — Resend創業者 Zeno Rocha 氏

It's like, oh my gosh, how are we ever going to beat SandGrid? It's just impossible. And then earlier this year, we crossed the number of downloads from SandGrid, and I was like, that's insane. And now we're more than double. So recent now is powering 4.3 million users as of this morning. We have all sorts of customers like Warner Brothers, HBO. There was a Japanese gang that recently came. they brought like maybe 20,000 domains that recent, like those are 20,000 different TLDs registered in different domain registrars. And we are a 10 million AR company. I would rather live in that reality than live in a reality where we're a hundred million dollar company. And my daughter doesn't want to play with me. It's not interested in me. And we're not that close. So we are, we're building the infrastructure and going like down to the metal now, but it's a big project. Like, we've been working for months now, and we're going to announce something soon.

Okay, we're in the studio with Zeno Rocha. Rocha? Rocha. Rocha from Resend. You and I go back to your VS Code theme days, where you're just an indie hacker. And I was like, man, I don't know what this guy does, but everything he does has taste. And then now you build an email infrastructure platform. Anyway, welcome. You're undersell just an indie hacker. Top 20 most active GitHub under 2022. Very, very active open source stuff. What's your typical how do you introduce yourself these days? Yeah, I think I just focus on my recent career for the longest time. Even today when I say like, oh, I created Dracula, people know Dracula more than recent. It's cool. Yeah, but it's been nice. Yeah, it's changing. Yeah. What was the impetus for, I guess, picking email? I think people always have, obviously, React email and all those things. I just don't know if I would have needed a company.

If I knew how hard email is, I probably wouldn't side the company either. Were you primarily JavaScript front-end? Totally. CSS? I was a front-end guy. No info, right? No HTML5 guy. And I think what was interesting and what drew me to it is I just look at all the tools around that, And I've used all of them before, the Mayo Guns, the Mayo Chimps, the Sandgrids of the world. And I've always felt like, man, there should be something where there's just someone should care about this space. That's how I felt. Like, you have all these players. It's a massive market, a lot of competitors. But no one that cares about the developer persona. Maybe they care about the marketer, the product manager, or whatever, but not for the developer. so I started just building around that premise of like if this was the stripe of email how would that look like um and yeah the like I had no idea or knew nothing about email like I was a user of email but I have never built like an email pipeline and and all that infrastructure before yeah I mean

like you know typically if you ask me like postmark was still a thing yeah like uh and then and obviously sendgrid and all these guys they're still around but like uh they're not maybe as innovative as someone who is like a new YC startup, like trying to make it. Just up front, I like to let people brag a bit. Can you, like for people who never heard of Resend, what are like the ways that you talk about like the most impressive numbers, logos, anything? Yeah. So Resend now is powering 4.3 million users as of this morning. We have all sorts of customers like Warner Brothers, HBO. So customers that are really big and have all sorts of demands, but also a lot of indie hackers and folks who are just building a website, building their own application. And, yeah, we just crossed 120,000 paid customers. Like the endless logo. Oh, my God. Yeah. You think this is effective? You want it to flex? There you go. You know what? Like in the beginning, I felt like I just need to have some sort of social proof.

And then in the end, I was like, you know what? This is actually pretty cool. And I honestly, like, I found companies like Warner Brothers just, like, on a random Saturday. I was just, this was before AI. So I'll go to my database. I'm like, oh, let me see who can I put on the website. And I just, I did the most naive thing you can do. Like, let me look at the domains. And let me just filter by .coms and filter by the small ones. And there was this domain, WBD, Warner Brothers Discovery. I didn't know. What is WBD.com? And I was like, oh, no, Warner Brothers is using us. So, yeah, I just love featuring customers. Specifically for listeners, one of the reasons I reached out for this interview, one, I should declare up front, I am an angel. When you sent me the memo, it was super simple. You were like, look, MailChimp has exited, Postmark has exited, SendGrid is under Twilio. Who is doing this? And I was like, okay, yeah, seems right.

Honestly, in terms of the pitches I get, I've invested in about 30 plus companies. That was one of the most straightforward things ever. And then secondly, I don't like to add API keys. I like to have everything on one platform, one API key to copy everything. So Cloudflare has built-in email. And I had a project, and I set up my Cloudflare email domain, DNS, all that junk, D-H-I-M, what have you. I don't know all the alphabet super stuff. Yeah, and I already had this set up. And then I was like, okay, well, I want to add another sort of email thing. Bloody model said, you must add resend. And I'm like, who the heck is this at resend? That like, you know, even already I have a Cloudflare email set up, and he's still recommending it. It's on the screen now. This is him praising and like, I don't need resend, but you want me to keep adding resend. Yeah, so okay, who did you bribe in OpenAI and Dropik to get resend on there? Oh, man, I wish there was someone to bribe because I'll definitely be paying them.

But I do personally spend a lot of time on that topic. Like, that's something that I feel like every generation, that's why I was so interested to build this thing in the first place. Like, every generation, you have those tools that define that moment, right? We remember the Heroku days. I mean, another 5% home. You know, and so to me, it was like, yeah, there's room for a new generation. And this new generation is not going to be distributed in the same places that it was before. So since the beginning, I was like very focused on, okay, we got to be number one when someone asks about that. You're very number one. Plug code recommending 60% plus of the time. I think it'll be fun to see the playbook. How do you win that? I do want to get back into, okay, so you said you had no idea why email, how to start email. I think a lot of people actually don't understand what goes into making a service like this, right? You might because you've set it up. Some people would just, you know, okay, I can use Twilio, I can use SendGrid, I can use you guys. But what goes into making a tool to just not send emails?

What do people not think about? Yeah, there's just a lot of different things you can do because landing on the inbox, especially when you're doing high volume, is different from like when you're scaling just like an app where you throw GPU at the problem or you just scale more vertically or horizontally, whatever it is. With email, you've got to respect the rules of the game and how Gmail behaves is differently from Outlook and differently from Yahoo Mail. And it's an industry that is fascinating because it's so old school and it relies a lot on contacts. But what we found early on was like, okay, we knew nothing about this world. And so in coming from this perspective, like extremely naive was the best thing that ever happened. Like, honestly, like it was the best thing because like there is no preconception of what is right, what is wrong. So for them, for example, every single app you go and then first thing you see is like this form that you've got to fill out about like how many emails are going to send?

What's the volume? What are you trying to do? And then you go for a human verification process. Two days later, they accept you or deny you. Most of the times they're going to deny you. And I was like, this doesn't make any sense, right? Like, why are we blocking people? So I know I'm going to remove that friction completely. And the friction will be on us. We're going to figure that out. Of course, now, four years, three years in into this business, I get it why they do that. There's just so much fraud and phishing and abuse across the board. I don't talk too much publicly about that, but we build like all sorts of crazy systems internally for that. But to your question, like how you architect that, like, for example, when you're sending an email to 100,000 people, if you just throw 100,000 emails like to Gmail, what Gmail will do is they're going to throttle the emails because it's way too much. And they want to see the engagement. So what you do is like they see like, OK, are people opening the emails? Are people archiving, deleting right away, or not interacting?

Okay, depending on those initial interactions, then what I'm going to do with the rest of the emails will dictate that beginning. So there are clever things you can do where 100,000 people put the most engaged audience members in the beginning of that queue. You have to put that in a queue. It's an art and a science where you've got to space out each batch So, and space out between the different providers too. And yeah, now we're like buying different IP blocks. You got to look at the history of the IP blocks, distributed folks into different IP pools, and then balance the IP pools over time. So it's a whole crazy infrastructure game. How automated can that be? Or is there a lot of human judgment still? Yeah, it's a lot of automation for sure. And I think over time we learned that, okay, this is how fast we can send for new customers. Of course, if you're sending more, then maybe you can send faster,

but there's a trial that we have to do to protect you as a sender. So what Gmail and all these players will want to see is a graph that is consistently sending more and more versus what most people do, which is you send ones to a big audience, and then you never send again, and then it's empty, and then boom, you send it again. Once a week, twice a week? Yeah, and those are patterns that an abuser will have, right? So the whole way of thinking about email marketing and email in general is you need to do exactly what an abuser wouldn't do. It's the opposite of that. So what does an abuser do? They buy a domain. We have now, like there was a Japanese gang that recently came. they brought like maybe 20,000 domains that recent, like those are 20,000 different TLDs registered in different domain registrars. And we look for every single factor there. So they come in, it's a domain that was bought two days ago.

So very new domains, they get bad reputation. Then you look at the website, is it returning 200 or basic things like that? Some people don't even set that up or they do like a lookalike domain. So they're like, oh, maybe I'll do another domain for code outreach. That's terrible because it looks like you're phishing, right? Like because you're using a domain that looks like the same. So all these little things, and they send a lot of emails all at once, and then they never send again. So you've kind of got to reverse engineer how phishing works and build a platform that does exactly the opposite of that. Do you have a list somewhere or do you recommend people go to some well-known lists? Yeah, what I've learned is, again, another thing I didn't know. Eno is the type of world where it requires a lot of education, but people are not interested in learning, right? Like the platform should do all those things for them. So we added like insights inside the product so people can just apply the insights.

So we look at setting patterns and now with AI is pretty cool too. Like we have link checkers and spell checkers, like all these checkers before you send. So, yeah, it's a mix of you're teaching your audience, but at the same time, the product needs to do the teaching. And the more you can bake in the best practices, the better. So one example, too, when you send an email, you can send an HTML version and a plain text version. Very early on, we were like, you know what? Like this, the whole world is changing. There are way more agents than before. And then before, let's just generate, auto-generate the plain text version of each email. We can extract from the HTML the markdown and then just send that. So we start sending that automatically, and that increased deliverability a ton. You know, it's so much easier for spam filters to just look at the content directly without having to parse that. So those things that we are like, oh, no, let's bake into that product. And most people don't do that, but we're like, oh, this makes sense. But with plain text, there's no tracking, right?

Because you can't embed the image. That's right. So you don't have open rates. Yeah. How do you know you have deliverability if you don't have open rates? Yeah, this is very interesting because the industry, there are two ways of tracking, right? You have open tracking and click tracking. Open tracking, what it does, it includes a one-by-one GIF image, a transparent image. And then if that image is loaded, then you detect that as open. The problem with that is that all the changes that Apple did recently in terms of transparency and data privacy, they read the emails before it lands, if they're doing the summaries and all those kinds of things. So the open rate as a metric, as a whole, isn't really as reliable as it used to be because all these spam filters are checking that before. So then you rely on click tracking. Now, click tracking can also be difficult because you want the URLs to match the sending domain.

You also want the images URLs to match the sending domain. That's another pattern that an abuser will do. They would host images somewhere else, point links somewhere else, and then they're sending from these domain links. I feel like I have used email services that would rewrite, like for the link tracking, right? They would rewrite to their own domain. You're saying that's bad. That's bad because what happened, we did that too at first. I thought that's normal. Yeah, but what's bad is when you have a noisy neighbor problem too, where if one of the customers using that click tracking is an abuser and sends a bunch of phishing, that domain might get listed on a blacklist, and then it takes maybe days or weeks to get out of the blacklist. Then you're screwed as a business. Is that not a pretty common? You make it very easy to just sign up, free tier, send a bunch of emails. I would assume you have a lot of spam, a lot of phishing, but you're able to avoid, and this is why. Yeah, I think what was different to me, and then this is more like a philosophy, right?

So I saw those other products and the way they behaved, and I felt like, okay, it's almost like if you were born in a society where you first go to prison, and then you have to spend like 18 years in prison. You've got to prove that you are a good citizen and then you get out of prison and now you can do the things you want to do. I always felt like, no, I want to live in a society where I'm born free and if I make a mistake, then I might be punished just a little bit or if I do something like really bad, a bad punishment, right? So that's how I feel about that too. Like you come in, we trust you by default. Of course, again, a naive way of thinking, But you know what? We're going to trust you by default versus make the whole platform worse for everyone because of those bad actors, right? So we trust you by default. But then what we do differently from others is like at the very first moment, like when we notice something is off, like by looking at the bounce rates, complaint rates, all the feedback we get from mailbox providers, then we block you like very quickly.

And then we chat with you and try to understand. And most of the time it's just misuse. It's not abuse. so we learned that we need to differentiate. Skill issue. Yeah, exactly. I think we have skill issues we wouldn't have expected. Yeah, email is so important and it's so dirty and it's so secretive and it's just such a I almost just want you to like just tell us everything. I will say you have a fun page literally your philosophy as you call it. I got lost just reading around this and then there's like a call, you have a pledge that you will make email safer if you're interested in this stuff you have quite a bit of yeah actually I do want to there's a how we scale support there you remind me of another investment I had, Jake from Railway there's also another very, very permissive, very develop-oriented but also you have to scale support because the moment someone gets blocked, they come to you and say hey, don't give you off your platform

how do you scale support? Man, I think if support was a software, we're probably on major version four or something. Okay. Like the very first version was like you, me, and I was so stressed out. I'm like, dude, I just need someone to be sitting on the chair. Like it's that simple. I cannot be doing this anymore. And it was like just anyone, anyone, I don't care. Then you do that. Of course, it doesn't go well because you realize very quickly you need someone that will actually try to prevent the tickets from happening in the first place. So then we went to V2, which was like a new version, like a new person that we ended up hiring that was like, okay, we're answering the tickets. They have domain expertise, but they were still like not doing the proactive work of like, no, let's make sure this is documented before we talk about. So one policy that we set, it was pretty cool, actually, is like you can only respond to a ticket with a link.

So you cannot just give the answer right away. You had to stop, write a knowledge-based article, write a docs. And then so it takes like an hour and a half maybe for you to answer a ticket. But guess what? Now you're sending that link all over again. So that was super cool, and it helped us build that knowledge base. But even so, like after that, when we crossed maybe 1 million users, so we had 1 million users back in December, and now we have four. So I can throw people at the problem, but that's just not going to solve it, right? Like I'm just going to keep staffing, and that's how it is? Like, no. So we really have to rethink, like, okay, everything from scratch again. And we want to keep, like, that free tier. I don't want to get rid of that free tier. I started as an engineer because I was able to use those free products, right? Like, we all, like, if it wasn't for Heroku's free tier and all these other products, like, we wouldn't be here. So I want to keep recent that way, but it took us many iterations to get to it.

How much of this is like, okay, it's a domain people don't know a lot about. There are actually unique questions that need human support versus, you know, command K with just, you have docs, you have a model. customer service plots exist, right? Yeah. I think right now is much better. There are a few things where, man, there's so much friction. DNS? Oh, my gosh. It's always DNS. It's always DNS, right? Like, the majority of the questions we get today is just, like, people not able to set that up. And it's fascinating because we're serving a new audience, right? It's an audience that doesn't even know how to troubleshoot. Like, I remember last year when Lovable came and Bolt and V0, I could tell these are different users. Like, they don't know when they have a problem, they are stuck. They don't know how to go past that. And with DNS, that was always the case of, like, yeah, they just simply don't know what is a C name, what is an A record, GXT, where do I put those things?

So they would paste the wrong value or silly things. By the way, right now I use computer use, like Codex or Cloud Code, and it just completely manages my DNS records. It's beautiful, right? It debugs it. It, like, digs, you know, and tries all the DNS stuff. Like, I don't have to worry about it. Yeah. One thing that we did, like, I found this super obscure protocol that, like, very few people in the industry were, like, investing on. It's called Domain Connect. And it's super interesting, by the way. And it something that started from some domain providers Oh yeah yeah yeah yeah This is like the OAuth for domains right Exactly Yeah this is like Cloudflare Yeah so some people have like Cloudflare is there Then we convinced Rysel to be there as well, and we started, like, getting more people. So you have the two sides, right? The DNS providers, the service providers. We are the service providers, for example, and then the DNS providers are implementing.

But it's a crazy, crazy industry. Like GoDaddy charges like just a lot of money for that. It's so weird. Like Jonas, it took us like the first one, it took us like months to get after that person too. Like no one really maintains the protocol and things like that, but it's beautiful. Like having a protocol that you can just do an OAuth for domains. Yeah. So for listeners, if you're not familiar, if you haven't tried it, something like that. It's literally like you do use Plaid or use like OAuth. you only click buttons, you don't type in anything, there's nothing to get wrong, you basically just sign in. As long as you trust this thing called Domain Connect, which you have confirmed that it is legit, then it's trust. Literally, I used it for the first time two weeks ago. I was like, what is this? I've never seen it before. So it's really cool. One of the things that I'm trying to figure out as well is how you start this company. Like you said, you tried to get in touch with the IONOS guy. I'm like, I wouldn't dare to start Resend without about having some guy on the inside at Gmail.

Even now, it's like, okay, can we vibe code our own email thing? Clearly, no, but that's even with tools today. You did this years ago. Like, you know, you want people at Domain Resolve? You need the guy, right? Yeah. You need to pick up the phone. I don't know. Yeah, there were many times that we had to pick up the phone. It's fascinating. Like, it's an industry that relies on relationship a lot. And we didn't have those relationships, So we hired for a few key roles, like deliverability roles, folks that have been at Postmark since the beginning or 15 years ago at MailChimp. And then we leveraged that network. But, yeah, it's pretty crazy. But what's interesting now, of course, I didn't know that when I started. Like I could never predict what was going to happen. You didn't do research? Yeah, no. Like I could never predict, like, okay, what AI was going to do and all that, right? But I think we are in a space where, man, I'm glad we chose infrastructure. I'm glad we're infrastructure as a service and not just, yeah, software as a service.

You had a little bit, right? You made React email the year before, so you did dip your toes in open work. Yeah. And then came the company. Yeah, React email is our next AS, right? So we wanted like an open source channel for top of the funnel. And I remember doing some things very intentionally. If you go to the docs, for example, you'll find there are docs about how do you set this up with SendGrit and MailChimp and MailGun. I wrote every single one of those docs over there. And I just wanted, like, okay, this is for everyone, and we're trying to redefine how this is. And it was a strategy of, like, we're going to give, give, give, and then we ask. And if people are... Gary V. Yeah, and if people are building the email template using React, now they'll need a platform to send it. Where do they do that? Here's Resend. So it was like a storytelling kind of way of approaching the market.

We're going to fix the first part of the journey, which is building the template, and then once you have the template, now that that problem is fixed, you need a platform to send it. Here's the platform. But did you have the idea for Resend when you started React email, or was it a side project that then became a company? We started Resend, and as we were building Resend, I was doing a lot of user research, and people were saying, oh, the pain of writing the template. I get the Figma file from my designer. It sucks. It's HTML tables and everything. They would cut the images, right? Yeah. Oh, my God. Because there's nothing. And then there's like a special syntax of HTML that only works on email. Yeah. Right? Or CSS. It's all inline CSS. Yeah. Inline CSS. That's right. And then like for Outlook, you have like some special comments and then you got to, it's such a mess. It's like a, it's a mess. And so to me, the reason that React works is that it is basically a compiler. Yeah. It's like, well, somebody writes this once and then as long as you fit the API, it's going to work everywhere.

Yeah. And you can import your tokens from Tayawang. Now you have a type definition with TypeScript. So no more, hey, empty, comma, you know, like on your emails. Like you can actually have like a string validation or whatever you want. I think one of the initial people that pushed back on React email is like, well, you're not actually using React. There's no state. And like React is supposed to be a version of state. So it's mostly JSX is templating. Is there actually like animations? Is there interactivity? No. Expanding accordions? No. No, so you cannot ship JavaScript on an email. So then what you do is, yeah, we basically use the templating. It's true and complete. Okay, so cool. I actually do want to continue on this email path, but I see so many names I used to know. I was just kind of wondering if you can indulge us a bit and just go through some of these names, what you learned from them, if you have any cool stories, just anything, you know?

Man, so much. Yeah. So much. I remember the very first year of recent, I was like, I got to meet the founder of Mailgun because I was still a user of Mailgun. And I talked to Ev. He built, oh, my gosh. Shruger? Medium? No, it's a different product. But anyways, like, it's a security product that he's building now. And I remember talking to him. I was like, man, why did you sell so quickly your company? What was the thing? And he told me like, oh, yeah, 90% of our R&D was justified abusers. So it's a terrible business. Like, you shouldn't go there. And I was like, I don't want to live in that world at all. So I was very worried about that. And it was a big lesson for me of like, okay, we got to automate this thing. Like there's no way we're just going to build like a massive org for, we do have a lot of people interested in safety now, but it's not like that's going to be our core business.

No, we're going to, we have to be pushing the industry forward. But that was one. And SendGrid has always been my favorite out of those ones and always been the behemoth, like the biggest, it's like, oh my gosh, how are we ever going to beat SendGrid? It's just impossible. And then earlier this year, we crossed the number of downloads from SendGrid, and I was like, that's insane, and now we're more than double. So in terms of unique users, what's the number? NPM downloads. NPM downloads. What's the SDK? Is that the best metric? Yeah. I think there's something about that metric in terms of adoption, but then if you look at how many emails they sent, it's obviously much more than us because they have the big enterprise customers yet and that kind of thing. I think one thing that's surprising is the whole point of the Twilio SendGrid acquisition is that you are no longer an email service. You're a communications platform, right? You integrate across everything.

It's not just about cross-sell. It is about I can reach you across multiple channels, something like that. Or that you have the size. You know, Twilio is like, I don't know what the size is now. Let's call it a $13, $15 billion company. You should have all the relationships, all the talents, all these things, right? So how do you compete with that? Yeah. Man, I admire Twilio so much. They kind of invented DevRel. Yeah, ask your developer. You know, ask your developer. And now when I go past that one-on-one billboard and I see customer AI or whatever, I'm pissed. I'm like, why? Like, you neglected the thing that got you there. and I think now Jeff is not there anymore the CFO he bought the onion yeah that's right that's right so I feel like they're just sleeping on it and first they did that with SendGrid and they're doing that with SMS too and yeah I think my lesson here

is I'm trying to figure out the economics of scale of this industry right like clearly you know let's call it they're 100 times larger than you when you just first started right yeah how come they cannot use their size to just win, right? So in other words, maybe there is no benefit to scale. Maybe you just make a better product and you can slowly. For sure. And I've always thought about, too, like, why did Twilio acquire Zengrid and not the other way around? Twilio was bigger. I mean, SMS. Yeah. It's big. But, like, if email and communication is a commodity, like SMS, your margins are even worse. Have you looked into SMS? Yeah, for sure. I think one day. What are the channels that matter? Email, SMS, what else? WhatsApp is big in many different countries. You don't pay anything. Yeah, but then you've got to go through meta and now business messages are going to be charged.

So there is some friction there for sure. Phone calls? phone calls is a good one uh i don't know i never take any robo calls right yeah that's true i'm less interested about that too what i care is like what are the things that makes me able to connect with my audience like that's what i ultimately care like i don't care about email dms yeah dms it's one it's hard because you cannot like dm a big crowd like but slack maybe is one right as I have so many Slack connects, and I need to share the news about this one new feature that I built. Yeah, Tylon. I don't know if you guys use Tylon. I think they kind of innovated this Slack support thing. But you can obviously trivially build your own. Okay, and then I think the other one I want to touch on is AWS SES. Yeah. So my connection to this is I learned to code with FreeCodeCamp. Nice. With Quincy Larson and everything. Yeah. They have like 5 million subscribers on the newsletter. Yeah. Sent through Amazon SES for like $5,000 a month, which is cheap.

Cheap. I don't know what you would cost. More than that. So basically, to me, it's like the low-cost player, right? Yeah, that is. I mean, it's Amazon. It's like they're super established. What do you take from them, and then what should people know to stay away from them? We started with SES, too. We knew that, okay, if we're going to be innovating here, I wanted to innovate on the DX layer because I felt like that was the underserved space. So let me have SDKs for every single language. I remember with Serval as we highlighted service because no one had guides on how to send any with Lambda or something. So there's just this underserved space in DX. So we started with the most high-level layer, and that served us really well. I was able to scale because we were using SCS behind the scenes. turns out there's just a lot of limitations when it comes to the way you structure domains

and they have this concept of configuration sets. And the hidden cost, if you're going with SES, is that, okay, you might be paying only 5K, for example, but now you need SNS for the events. So you need all the feedback from bounces, complaints, open clicks, all that. And now, so you get the events, great. Now you need CloudWatch for the logs. Now you need a queue because it's going to be a lot of events. So now you add SQS. And now that you have SQS, so you start, like, adding all these different layers. Oh, you want webhooks. Okay, now you've got to build another thing. So if you go with SES, it's a great platform. It's fine. But you're going to notice, like, your footprint will be much bigger than you would imagine very quickly because you need all the apparatus around it. Yeah, I would say, you know, another YC startup that came up in the mail areas, Asian mail. And then I talked to the guy as well. And he was like, yeah, we're going to build on top of YC.

Yes, we're not going to compete with them. Yeah. But for us, it's changing that now. So we're building the infrastructure and going down to the metal now. And for us, it feels like we got to the point where now this makes sense to innovate on the most basic layer. And now we can get better throughputs, better deliverability if we just focus on that. But it's a big project. Like, we've been working for months now, and we're going to announce something soon about that. What is the hardest thing to – is it, like, trusted IPs? Is it compute and data centers? Or what is the blocker? Yeah, definitely IPs. You're buying, like, a whole IP block, and then now you're managing the IP pools yourself. That's one. and then you're almost doing like traffic control and you're controlling the throughput into each mailbox provider and then you need the relationships with the providers. But then like once you have that established, that infrastructure now, you have way more control into what you can give to users.

So one of my biggest things were like observability, you know, like as an engineer, I want to see exactly what's happening and I've always felt like in this space, there wasn't enough information. Like when an email goes out, is it really delivered? Is it really on the spam folder? Is it in the promotional staff? Like I never know. Like you just trust that the email goes out and then you always hear from some users like, oh, I didn't get the email. I'm like, what? Like what happened? So there's that empty space between those things. And the more we're close to the metal, the more we have in terms of data that we can surface for users. So I'm really excited. It's like the flagship change that we're making now, too. I mean, you're becoming, it's really interesting for all DX type businesses. You start with a thin layer, like a React email, and then it's like, okay, well, we'll build on other people's infra. And then once you get size, then you can actually build your own infra, right? Yeah. Are you doing anything differently because now you own everything? Yeah, I think what we're doing is...

And by the way, obviously, your dashboards are beautiful. It's funny, man. And everyone I talk to now, like all the customers are like, oh, I never opened the dashboard. But we can talk about that later. But it's interesting. Like I spent so much time trying to make that dashboard beautiful and now people don't go there. It's kind of frustrating, but I get it. But yeah, we're just trying to approach that with the same naive mentality that we approach all the other problems. And we're like, okay, this shouldn't be this way. So you're challenging the most basic things of like, this shouldn't be this way. The P99 shouldn't take this long. how we can make it faster. Okay, so that's one of the things we're doing now, like focus very, very deep into performance. And we have always been like a Node.js shop and we have like just TypeScript everywhere. And most recently, like the sending path, we no longer use JavaScript, we have Go there. And we're changing other parts to be more performative. Yeah, because now languages don't matter. Exactly, right?

It's just so cheap to, yeah, like pick a problem and yeah, you can rewrite the whole thing in whatever different languages. I'm like running Rust projects and I've never, I don't really know Rust. Like I know like some theoretical things about Rust but I've never really bothered to learn it. I feel like we are not really covering the recent products of this area. Maybe you want to go. No, no, I mean, you know, there's everything we've talked about, which is why it's hard, what all you're solving. And then there's also just send emails from CLI, super agent forward. You have some takes on this. I think the other day you tweeted, like, MCP is still back, right? Like, over a million MCP calls. Like, what's the AI side of email? Yeah, everyone was saying MCP was dead, right, a couple months ago. Like, yeah, the only thing that matters is the CLI. What I've noticed is that I stopped paying attention to what's in, what's out, and I really think about these problems as like, okay, how do I want my product to be? well, I want my product to be formless.

Whatever they want to do with the product, it should be possible. So that's why we've always been API first. I had no idea how much this would help us down the line, but I knew, yeah, I will never be able to cover all the use cases. So let me just give the API, you do whatever you want. And with the MCP, it's just fascinating. Every month, at least 2x growth, and now in August, these numbers are even more. This was before the end of August. I think it was 1.2 million by the time it ended. But, yeah, people are starting to work from cloud desktop or codex. And if work starts from there, then you need those integrations, right? It's just super important. I just want to clarify as well. So, you know, you showed us sort of like the month-by-month numbers on the MTP thing. It picked up from, like, let's say 50% month-to-month growth, which is already super high. Then it starts to double basically every month. Did you do investment into MCP or you just had an MCP and it just, after some time, it gets picked up?

Well, fun fact about that, our very first MCP was built by Yoko from A16Z, our investor. Wow, value add. Yeah. So she built the open source version. It was a local MCP. And then it started to grow like crazy. I was like, man, why are people using it? It's like so hard to set up MCP. This was last year. And then, Yoko. Yeah, exactly. So she built that. And then I just asked her, hey, can you transfer to our org? It was just like an engineer building something. And it was nice to see after the traction it was getting. But to do the MCP right, to do the remote one, we had to go back and revisit our entire permissions model because we needed OAuth. And then we were like, okay, now we have a blocker. Let's build OAuth. So then we build OAuth and it's like, okay, our back, let's look at that. Let's look at how users have access to each one of the resources.

So it was a whole journey to build OAuth. And then after we had OAuth, okay, now we have the remote MCP. And then it's distribution. So then, okay, now we have the remote MCP. okay great great great observability is in place great great okay all the data dog dashboards okay now all in let's talk to everyone and that was hard too like uh it's not as easy i wish i had all your connections uh it's not as easy but like talking to the labs and trying to get into like the the official stores and plug-in marketplace you think that actually matters oh for sure Yeah, so I think sometimes I've been on the other side where I try so hard to get these partnerships, we launch it and then nothing. All right. So which one matters? So as a consumer, I make decisions now based on do I have the access to that data or not. So as a buyer of B2B software, if I go to the cloud connectors and I don't see the company listed there,

with an official MCP. Like, we had this. We were using Plausible and Pulse Hog. Plausible had an MCP built by Sentry. I don't care. If the Plausible doesn't maintain that and that's not official, I cannot trust that as much. Even though I love the Sentry team and I trust that team, but I need the owner of the software to own this piece of integration. Oh, my God. And I made the call like, no, we are getting rid of Plausible and we're just going to use PostHog here because I have the MCP there. And I did that as a B2C consumer too. I'm like into running and then I'm like, oh, let me like Strava, Strava has an MCP. I can like access my runs. So I use Strava because of that. So, and if there's like, oh no, we have an MCP, but you have to connect like this, like the X MCP, for example, I have to do like the local one. It's just so frustrating. it works on the desktop line, not on my phone. Like, to me, it's just, yeah, if it's not there,

I don't use it. Apart from, okay, it is officially maintained by the vendor or whatever, what makes a good MCP versus a social MCP? Even for hours as I interact with that there were I always like it always collapsed right The the tool calls I always opening and trying to understand like what did it do And just a couple of days ago, like, there was one call where it couldn't put everything on the contacts for some reason. And then I'm like, no, this shouldn't happen. Like, how much data are we sending? Maybe we're sending way too much data to the client. So that's something that I noticed, like, the amount of tools, This was a bigger concern earlier. Like, oh, if you have way too many tools, they get confused. I don't see that as much anymore. And, in fact, we're just adding more to the NCP. Oh. Like, we had, like, maybe, I don't know, 20 tools when we launched or even less than that. Now we have, like, 80-something tools. What the fuck? Yeah. And, like, I'm adding more and more.

Like, the long tail of, like, the footprint. That's what I've noticed. You can do only 90% of what your dashboard does, and that's one way of doing it, or you can do 100%. Getting to 100% coverage of what your dashboard can do is actually very, very difficult because now you're doing usage, billing, team invites, account provisioning, deletion. Should that be exposed as an API or not? It's really hard. It's a difficult problem. Yeah, in fact, you are one of the few Linux-based guests that I can ever talk about this with. I think this is one of the mistakes of Next.js. Oh, really? Because Next.js and server components and Pages API or Apps API, they make it easy for you to basically bind data to your page. So you never expose a public API. Yeah. So by default, the thing that is performance and they encourage you to do, you will not have an API, right?

Then you have to just do separate code paths or whatever. In fact, you should just always have completely headless backend. And then your frontend just happens to be one of many clients that connects to your backend. And we started with our API on Next. I remember that. First of all, the view was just so high that we were like, there's no way we can pay for this. So then we moved to a traditional node server, server four, whatever you call it. and then like our dashboard was always like a private API consuming the same thing. We had to duplicate a lot of code. It was a mess, right? Now it's so much better. Everything on a monorepo, you can reuse stuff, but yeah, like by default, ideally those frameworks like just gives us those components, right? We had to build a public API separately from that and it sucks to maintain both. Yeah. Okay, and then one more thing about just a PCP part So you've got 80 tools in there.

You're growing more tools. You are very counter consensus compared to Cloudflare, which only has two tools. You search and execute or something like that. So far, you haven't felt that issue. But what would make you change your mind? It seems like numbers are going up, but do you actually know? Do you talk to your customers and say they're like, oh, my God, it's like the best NCP on earth? I do try to get a lot of feedback on that. and then the way I like to build products is I've got to use the product on my own, right? So if I'm like the investor updates that I sent you yesterday, that was built using the MCP, you know? It was like fetching data from one place, from another place, and then... You're making me feel bad now. I don't do all this. I don't use MCP now, basically, is what I'm hearing. I mean, I guess the other thing is also like you do have CLI, you do have regular SDK. Okay, like, are those also just going crazy? Is everything just from a lot of growth or? Yeah, the CLI is definitely growing, but not as much as the MCP.

But I have a lot of fun with the CLI too. Like, if you go to our careers page, I'm actually proud of this because I did it on a weekend or something. But if you go to any job there and you try to apply for a job, and you have bills to go down now, you have three ways of applying. You can apply to the CLI or API. So you can just log in and it has an interactive mode or we have skills baked into the CLI and you can just ask Claude, like apply for a product designer role at recent. Okay, so you are filling yourself with slop. Especially in careers? Yeah, yeah, yeah. No, I mean, I think maybe that's also the people you want, right? Totally. I actually have a filter. if the source is the CLI, I look differently. Because I know people are going the extra mile, like setting that up like, oh, no, I got to download this thing and I have to authenticate and then I have to go through that flow. But guess what? Like people who go through that, they always tell me like,

man, I've never applied for a job using a CLI before. So even if they are not accepted, like there's just something, like it's just a different experience than other job applications. Even API applied for one job during the other. I mean, it's the same thing as JLI. Yeah, so I think one of the things that I'm trying to figure out as well is product philosophy, right? Like, you're clearly just max everywhere. Every SDK, every language, every surface area. Obviously, easier with agents, but I think some other founders that I would know would say, hey, this doesn't matter. Give that to Greenhouse or Ashby. Yep. It's your job application. You're not going to differentiate like this, right? Yeah. You clearly have a very different philosophy of you care about everything, even though you know probably some of this stuff is just wasted. The way I see it is I'm trying to optimize not only for the privileged users, like the users who have the greatest Wi-Fi, the MacBooks that are here in SF or in New York, right?

all these people in the margins that are in other countries with a terrible Wi-Fi. Brazil. Brazil. Growing up, that's how it was for me. A terrible Windows computer and terrible Internet access. So that's why we have the free tier. We don't have the free tier because there's a reason behind it. When we have all the SDKs, I'm maxing on the SDKs. man, building the .NET SDK, do you think that was fun? And we looked into... Stainless. Stainless, Fern, what's the third one? But we looked at all these companies, right? We didn't like what was generated. So every one of those SDKs is humanly generated. The Java one, the Rust, all those ones, right? And I do that because I feel like, yeah, if you're using a product, You should have the best experience in that language that you're using. Oh, but there's only 500 people out of the 4 million.

I care about those 400 people. Wow. You know? Why not, you know? Why not is you have 4 million users. That's why not. There's a lot of stuff to do, but, you know. It's good to have that opinionated side towards building this, right? Any other product surface area stuff that we should cover? You covered CLI, covered templates, covered MCP. I think it's good to get a high-level perspective from you of how much of this is very traditional. You even have dashboards, templates to manually write emails. You have programmatic. You also have the full agent build. Where do you see it going? How do you think about building this for a pretty fast-changing world, right? Yeah. There's definitely still a component of email moving out of this single-player mode and going more into multiplayer. We even built our editor like this. It's multiplayer. It fixes all the conflicts if you have two or three people. We actually run, I have a private recent account with all the teams,

and we run all of our all-hands there. So the all-hands is pretty much, you go there, it's a huge. You make an email. It's an email. And we send the email together at the end of the all-hands. It's super fun. And you have like 20 people in the same document. I feel like 99% of people have not thought about multiplayer email, right? But yes, you have a lot. Yeah. And it was really fun. Man, that editor alone, it's so freaking hard to maintain that thing. But I love it. I love that we have the UI for people who need the UI. Like some teams hate to have the email template pushed into their source code. then they just want to build a visual editor, a template, give to the marketing team, and just call the ID. I don't care about all the templates. I just want to call the ID with variables, and I'm done. Other teams, they do like, no, I want to commit. I want to have source control over this and own this piece of thing. So I try to tailor to those different audiences and build a product that feels that we care,

which is not the feeling I get elsewhere. Yeah. By the way, what does the recent team used to read email? Superhuman? Yeah. Oh, man, have you seen the last Superhuman update this morning? You were busy with other launches too, right? Yeah, and for listeners, we are recording this on the day of the Astro launch, which I was pulled into, and I am not sleeping. But what's wrong with it? They just updated the UI a little bit. Yeah, there's just a sentiment of like, oh, this is going to the direction of every other acquisition goes, right? Like now you're fitting a brand that has all the... Grammarly and other things. Yeah. Okay. It's just email. And I do want to keep up to date on email tech in general. You're on the sending side, other people are on the reading and receiving side. Yes. And then I think, you know, increasingly I would say agents reading my email. Yeah. I actually don't read most of my email. I have Polk and two humans helping me with my email.

Nice. I love that. Polk is pretty good. Honestly, like Instinct now super hot, $2.5 billion company with no revenue. And like it's just like people trying to read email evasions. Yeah. We are definitely looking to that space. That's another thing that I'm spending a lot of my time, like how you can programmatically create dozens of inboxes and then how are those inboxes accessed by email? Because you enable receiving email. What is it called? Inbound? Inbound, yeah. Email? So now we're going another abstraction. Inbound is like the most basic abstraction. Now we're going one step further and be like, oh, you need 100 inboxes for any reason, and here are the APIs that you need on top of that, right? So you need a read API, markers and read. You need star API, labeling. I see, I see. So we're going to be launching a product around that soon. Yeah, but again, it's all API-driven because that's the... Fully, yeah.

Because I would say the comparable stuff that I've been looking at is Missive. Yep. And there's other, like, Help Scout? Yes. Like a shared inbox type of product, right? Front. Front. Yeah, so hopefully this can be a version of that. I can share with you. There's a feature flag for you if you want to try that out. Yeah, I'm very interested in email tech. But okay, and then to me, again, I'm just asking all these questions. What percent of your email goes to a G Suite or, you know, like what's the market share, right? Like Google versus Outlook versus Apple versus Hotmail. Yeah. I don't have the numbers top of my head, but the vast majority, like at least 60% Gmail. And then you have like maybe the big tree after like the Outlook, the Yahoo mails of the world. right after, and then you have a long tail. Apple is not that big. Apple is big, yeah. And especially with the privacy thing that they did now,

it's harder to sign up to an Apple. What? Yeah, like Apple auto-generates an email address for you. Oh, yeah, yeah, yeah. Yeah, I use that. Is that a problem for you guys? They're more strict in terms of what can get into those inboxes. so sometimes as a domain owner we actually have like an in-product wizard to help you like register with Apple so that you can get your emails delivered there. Yeah, yeah. Okay, I understand. Obviously this is your job, right? The whole thing is deliverability. If you don't have deliverability, you don't have a business. Exactly. Yeah, that's kind of the... Yeah, we can have the most beautiful Rubik's Cube in the website. The most coolest handbook ever if it's down. Yeah. Tough enough. And how correlated are they? If you get banned or downranked or put in promotions tab in Google, does that transfer over to the other guys or are they very decentralized? Thankfully now, like in the beginning of the email industry, it was very IP-based.

And if one IP pool was damaged, like now you're getting the consequences of that. Now there's a lot of reputation associated with the domain. So then what people do is you create different subdomains depending on what type of emails you're sending. Like one subdomain for transactional emails that you know that maybe doesn't have click and open tracking. And then you have another domain for your marketing emails. And those subdomains help segregate the reputation of whatever. So would it be like transactions.resend.com? Yeah, updates, notifications, whatever you want to use it. And do I have more dots if I want, like a.b.c.recend.com? You could do, but we just see like doing, I don't know, five subdomains is like, yeah, not really practical. But like people do, yeah, like some special ones. Like we do that at recent for our own emails. Like marketing has a different stream than the transactional ones.

And yeah, even the investor one has a different, like I use confidential.recend.com to send those emails. I didn't even know you could. And that is good advice. It's a good practice. Before it was very IP-based. Now it's very domain-based. You know, I have received so many bad advice. I have to buy a separate. Because we use AI.engineer and latent.space. These are all weird TLDs and nobody recognizes this, right? So I was told to get .coms for everything. And then maybe that improves deliverability. The problem with that is that now, let's take Coca-Cola, right? Coca-Cola has a domain. Now, you're Coca-Cola and you're like, oh, yeah, let's buy another domain that looks like our domain. It looks like fishing. It looks exactly like fishing. So the chances of damage is way worse. Is this still a thing when you're buying the .com? I get it the other way, right? You're latentspace.com. I buy latent .space. It looks like fishing, right? Yeah. The .com still? Man, you would be surprised.

It's like there's just like the things I see on a day-to-day basis is like crazy. Yeah. So what I'm hearing is like, don't come, it's worth it. Yeah. That's all the sort of business of email. Obviously, some part of it is you just got started before ChatGPT. Yeah. And then you were a leading player that people liked. And then obviously organically you pop up. You didn't do anything special for AEO. What did you do for AEO? I would say we were not the leaders because we coming up like in 2023, like very few people knew about Resend. And we were like so behind. Twilio, Postmark, Mailgun, those companies invested in SEO for 10 plus years. So for us to get ranked for anything was like almost like it was an uphill battle, like so, so difficult. So then what we did was like, okay, let's focus on the docs for sure. And the docs was very intentional about like, okay, if we're doing a docs on,

it has long tail again, like, oh, here's a docs on not just Python, like how you set it up with Django. How to blah with blah. Exactly. How to blah with blah. Super short articles. This is traditional SEO. Traditional SEO, nothing fancy there. And then when models started to come, then I was like, okay, maybe I've got to change the content of the page. Because if people are asking questions, then I need to create content that is in action and ask and answer type of format. So a lot of FAQs. Now, almost every page at RISC has an FAQ. Like, you go to pricing page. Of course, pricing almost always has an FAQ. It's pretty standard. But you do FAQs in all sorts of... Yeah, I think maybe security also has, yeah. So you see, it's like a pattern of like all pages have that because that's how people are asking. Is Resend Socks to compliant? And then just have the answer right there.

And then, okay, now we have those things in, okay, now we got to do llms.txt. So we added llms.txt on our docs, thankfully because of Mintlify that we use, great product. And we're like, okay, how we take this to the marketing site too? So then LLM is a TXT for every marketing page. Same with the docs hatch, markdown versions of the docs. I'm going to do markdown versions of every single page on the website, on the marketing website. You think content sniffing or you just append like .md? Yeah, just append. You do append. So if you go to the security.md, it will work. But also if you do content negotiation with that page, it will render on that page, like the markdown version, without going to that. Yeah, which for listeners, this is one of those things, I'm not sure it matters. I'm really not sure. Yeah, I mean, you're winning really well. So, yeah, I think one example that changed,

and maybe it doesn't help with distribution, but it helps with accuracy. So go to the pricing page. I remember this was like a random Sunday, and I was just like doing competitor analysis. And I was just asking, like, okay, if you send 1.2 million emails using Resend, and you do that with X and Y and Z competitor, who is the cheapest and who is the most expensive? And it got completely wrong. It got the Resend one, like, so bad. And it actually said, like, very bad things about us. Like, oh, Resend's pricing is very obscure. I was like, what the fuck? Like, it's not obscure at all. So it was, like, super frustrating. Oh, so you made a calculator. Yeah, so the problem was, you know, agents don't really drag a JavaScript widget. Okay. So then what I did was the .md version of this page, which has a tremendous amount of access, by the way. Like now, like...

I see, you get analytics on this. Yeah. Yeah, yeah, okay. Not through the regular analytics spots, too, because, like, we cannot track that. Yeah, it's the server side. So then, like, we got to go to Datadog to figure that out. Oh, okay. And, man, it was really crazy. And then I made that change. I pushed that, and then I asked again, like, a few days later, and then it got the numbers right. I see. So I'm like, okay, this already – It's a good way to live update and fix stuff. Exactly. And this is what matters, right? Like, okay – But it does mean that you don't have a single source of truth. Like, you have to update the MD, and then you also have to update the – That sucks completely. But again, if I wanted the happy path and the easy path, I wouldn't be doing email. So I'm like, that's the price of like... Well, you know, if you're good at React, you can write a custom renderer. You know how to do this. What else? We talk baselines, basically have FAQ everywhere, basic SEO stuff. Comparison pages? Us versus SendGrid, us versus...

Although our page could be much better. Go to... Yeah, notice is not there. I mean I'm sure you have it somewhere in some footer or something or whatever just do slash migrate by the way 404 pages I always think we should guess pick the slug and there's neighbors oh yeah okay in case you ever break a URL it'll just autofix itself you don't just have the 404 my intention with those is like oh let me try to do something that's not exactly like a comparison page But I don't think it was a great idea because our competitors actually built the comparison pages and they get recommended more than ours. But what we try to do, like if you go to Postmark, for example, is like, okay, let's demonstrate. First, like, hey, if you want to migrate, here's a code converter. Is that an example? Yeah. So if you just put any code block there. Ooh. Why?

Why, dude? Why? that what we do that what we do this is no extra There a cube on the landing page Everything is extra Here a lot of continues But, like, the adoption is one that I pushed this last weekend. Like, now we have a graph that compares, like, our adoption with every different competitor. And you see one thing. Like, again, this was the lesson that I learned with the pricing page. JavaScript widgets. They're not going to be executed. So I spell out the same data in text so that LLMs can actually read that thing. I think they're getting somewhat better. I see a lot of agentic use. Okay, load website, click on image, interpret image. It's slower, but with computer use, they are getting much better. I agree. But if you're fetching information and you're just

comparing stuff, they don't do that range. I think they're starting to go more towards computer use. I've been doing a lot of compare this and this and the browse of website is pretty computer use. I noticed that you don't have another classic AEO tactic is like, here's how to connect our thing with clock code. Here's how to connect our thing with codex. The whole eyepage. Oh, okay. There are some things that are interesting, like if you go to the legal page. So, recent.com slash legal. Any one of these like, oh, you need to summarize or asterisk plain. Complexity. I see, that's just an external link. Yeah, yeah, yeah. Yeah, again, not sure that matters. You mentioned a lot about recency for AEO compared to traditional SEO, so. Yeah, that was my the thing that I was super curious about in the beginning. Like, there was no reason for any of the models to recommend us because we're so behind in SEO.

We don't have the typical glossary page or whatever, like, or articles that are so generic that we're just, like, SEO heavy. And I was trying to uncover, like, what is that? And I could totally tell, like, at least with Opus 4.5, there was a recency bias in terms of the technology and where they were fetching data. I know recently, like, they said, like, Maybe Rap Lead is not necessarily something that people suggest anymore, but definitely newer tools, they started popping more, and Reason was one of those. And because there was a lot of, I don't know, organic content about the tool, then we would be getting recommended more. So the advice here is just try to get as more people in more places to talk about the product as possible because it definitely influences. Obviously, I think that is a class of it. I guess, did you ever engage SEO consultants, AEO consultants? Are they useful? Are they waste of money?

We never really did that. You seem like you're just doing everything yourself. You're the AEO. You have a 2024 tweet of let's talk about AI SEO. I think you're one of the first ones to pick it up, you know. I mean, I think people generally, the state of things, because we did an AEO or GEO episode at the time around this stuff, and most people were like, just do what normal SEO does. And don't block AI chatbots. I think most tools like Profile at the time, they would just track your mentions and then let you do experiments and all those things. All those things are par for the course and normal. I'm just trying to, obviously, people are pinging me because I'm also doing AO experiments on what to do. I have unfair advantage because I'm a content business. So obviously, I pump up content, whereas you actually have a product. how important is it to you too like you're very winning but how much do you keep a track of how well are docs used how often are we recommended compared to competitors like

man i think it's so crucial because before we if we want if we're like livid and sure about which tool to use we would ask google and then you would get to like those comparison pages or you get you like the first three responses and now you just get one response out of the gate right if you're asking which two should i use and yeah people just want to go with the the one that is the most frictionless and yeah which i think free tier is very important and um i think uh michael grinich from work os is trying to promote this idea of off md right i don't know if you've seen it so so uh the idea that you it's very similar to netify as well where you don't need an account to do the basic action. So it's like unauthenticated usage of your product, which is very dangerous, obviously, because it's prone to abuse, but it is the easiest, like no sign up. Yeah, that's the thing that I'm trying. Yeah, I'm debating doing that or not. Do we need a provisioning API?

And I keep going back and forth on this, like, okay, call this API. Provisioning and claims, right? Yeah, provisioning and claims. but then free to yeah like i don't know if it's something about our industry but yeah you just have abuse yeah yeah just have a lot of abuse yeah okay yeah i mean i i think those are the the main tactics i i think the other thing that uh i would just tell listeners to think about is literally hook your clanker up to your analytics and your code base right like you need to like fully complete the loop and then like basically auto research right like like this like so what I have now is my cycle. I have both I have a latent space CMO agent and I have a AIE CMO agent. And they spawn other agents. Their job is to research ideas, keep track of things that we've tried. And then I sign off on things that we want to experiment. We have an experiment window. We go and do the experiments. Oh, really? And then those are other agents that are monitored by them. So I would say

that's kind of like the auto-research loop for doing this stuff. I think then you also need to give it tools. So I signed up for Ahrefs. I signed up for data for SEO. And then most recently I made my own AEO tracker, which you can make. And then you just want to make the number go up. For sure. And it's like, you don't know, right? You're throwing things at the wall. That's why the thing on the legal page, does that matter? I have no idea. But I'll keep throwing things at the wall and seeing what... I think Reddit, doesn't matter. Hacker News. You used to be big on Hacker News. No less. Not anymore. Twitter doesn't matter as much. Yeah. It's weird, right? It's so crazy. Do you have a tracker? No, like, we use all of those Profounds of the world. There are other products. So we tried them all. Like, people ask me, like, oh, should I use Gauge or Profound or this other one? I'm like, yeah, we have four subscriptions. You know, like, that's what it takes now. Like, you've got to see which one works best.

What I miss from those tools is, like, give me the action. And I think that the problem, they don't give you what to do because they're also trying to figure out what needs to happen, right? Like, oh, write an article on Wikipedia. Yeah. Yeah, I would say the old school SEO tools like Moz and all those kinds, they actually have more opinionation around these things. Right now, the SEO tools mostly, from what I hear, again, I'm not an expert, we should probably talk to profound, is that they just are reporting tools. Yeah, they're analytics reporting tools. And that's fine. Yeah. It's a big business. Ahrefs is a Singapore business and apparently they do like, I don't know, three, four hundred million dollars. Yeah, totally. The other parts of your business, we're going to talk about the parenting side, the personal side. I just really want to be like, okay, you seem very chill. You have 50 people remote all around the world. You are worrying about MCP, you're worrying about AEO. That cannot be the way you're closing all these big customers, right? Like, where's your fancy sales guy

who just ate dinner and a golf course? we're just starting that motion yeah and i think we are uh very lucky that we're able to go this far without having that motion i've worked at a enterprise only company in the past so i knew what does that take and the steak dinners like all the the apparatus around like that sales org enablement there's all these things that are needed i knew that okay when when the time is right we're gonna have to turn on that key but i've been very cautious about doing that a reason like in different layers instead of just like add 20 i'm very maybe it's because we are remote but i think we are very protective about the culture like let's make sure we don't damage the culture at once so no let's try to figure out people who fit so we just have one a now and it's been helping it's been yeah going super long yeah good enough i would say i think you know once you have

like you know you double your ar because of like a couple deals you're like okay like this is a good yeah because people want to pay you more yeah for things that you're like okay well you know i have like the agreement with you i have like your software on my premises or in my vpc whatever right yeah um so for sure yeah and it what i'm learning too is like you're building a different product. So that's how you need to see. It's not only I'm building a different machine internally as a company, as I'm designing the company, I'm designing, okay, which department who sits where, how each person is better equipped to do their job. No, like for an enterprise, it's actually a different product. Like what do they care is very different from what $20 users scare. And if I just add, you know, work OS and SSO and scam and audit logs, it's not going to be enough. So it's not just putting those items on the pricing page. So it's a journey for sure.

Any other things you want to reflect on the journey before we go into the personal side? No, I think we covered a lot. Okay, then I'll talk a little bit about my use case, right? I'm a content creator. Nice. Right? The bulk of your business, I assume, is transactional email. I probably could look it up. That's right. You have marketing emails. You have inbound emails. I'm sure that's sort of less outside of the business. I pay Substack something like $10,000 a year to send my emails, and that seems too high. Yeah. So I think, like, why is marketing email versus transactional email so different? It's a great question. Because your business model is also different, right? It is. It is. And we are more expensive. Yeah. Yeah. Then I was like, I can't do this. Yeah. The industry has this pattern of like charging per contact. That's the first thing. So the way they try to price is around the value you create in each contact. But now it's crazy.

Like the AI companies that we serve, they grow so fast. Like Mistro uses recent. Okay. So they're like the amount of users that they add is just insane. So like the business model of contact-based pricing is kind of broken, and we are realizing that and trying to fix that too. But that's how it had always been. And there's something about like when you send a transactional email, it's just like put a message in a queue and get out of the queue. With marketing email, you're building that engine of like batches and intervals between batches, analyzing, increasing the speed, decreasing the speed as the sending is going, maybe blocking. Is there a max rate you should send emails at a time? Yeah, for sure. Or domain, specifically. So let's say I have 100k emails to send out. It should take me 10 minutes? Yeah. And for example, now with the 4.3 million users that we have, it takes more than a day. Oh, what? No, that's unacceptable.

Yeah, for sure. So then I pay you more to get in a different queue. Yeah. So you have to be always trying to figure out, okay, we need more at the door, so how does it work? Yeah, yeah, yeah. Oh, that's interesting. It's like read limits for the state token. Yeah, exactly. Where's my reset button? Yeah, exactly. You also, like, you went into YC, you started this company as a young father. And I think, like, part of the thing is, like, there are a lot of no-kids people in SF, and I think that's part of the consideration. I think a lot of people are listening who are like, I can't be a founder, like, I have this great idea, but like I got young family. How do you manage? You got a preface with you left SF for kids. Yeah, so this is the recent news. Like I'm spending more time in Brazil now because we're expecting our second kid. And I think at first, man, it was crazy because I came to SF, my daughter was one year old and every WIC founder around me had no idea

like, yeah, what does it mean to have kids? and what is this reproduction you don't you don't even have series b like what are you doing married like totally most of them weren't married at all and it's crazy how um i think it again i always try to look at those things at first you you see them as a the advantage like oh man all of these guys they went from from like they came from mit all these amazing schools that was not the case for me. Oh, they don't have, they are not married. That's not the case for me. They don't have kids. That's not the case for me. But still going for that program, I'm like, man, I just want to build the best company on this batch. Not number four, not number five, like number one out of the 250. Oh yeah. Which cohort? Winter 2023. Okay. You realize in the middle of that is that there's no winner in YC. There's no such thing. uh but like i'm pulling a winter 23 now let's see who i know yeah and especially what you

understand is that uh winter 2023 it's like an infinite infinite game right like those even if you had an amazing batch and you raised the biggest round it doesn't matter because like life goes on and it continues and many of the people that i met there they are not here anymore Yeah. Co-founder breakups, seven pivots, and they're dead, right? But we are here. So I was very much like, no, no, no. Even the domain, when we bought recent.com, I'm like, no, it's going to be an amazing website with an amazing .com. It's not going to be a .io or whatever, right? But doing that with kids was so freaking hard, especially because I wanted to enjoy my first daughter. And I really thought about this. Like there was one point during YC. So this story like translates to what I'm going to say. I came to this event and PG was there. So I'm like, oh man, Paul Graham, the father of everything, God is here, right?

So everyone is paying attention to PG. And then he says, you know what? We always tell you, don't do this thing. We tell you that as partners, and then you end up doing the thing, and then months later you come to us and say, oh, I regret doing this or not doing this. So he just said that. So then I came to my partners, and I asked, like, okay, what are the things that I'm doing now that I shouldn't be doing that you think it's wrong? And then one partner, not my main partner, but another partner told me, like, you just worry too much about design. You care too much about that. If you didn't care about that, you will be much further ahead. and that hurt me so bad because it was true like my slides when I was like pitching it was like a beautiful slides and she's like doesn't matter I'm like yeah now I know it doesn't matter but like that hurt really bad and then a few days later I was reflecting on that I'm like yeah but if I if I were to build a product that had like this identity and only 1,000 people used I prefer doing

that than having a product with 10,000 people, that it's a product that I don't want to go, that I don't use myself. So you know what? I'm going to follow what I believe is right. And I think in this industry, again, going back to the industry, everyone had terrible products. There was not enough design infusing to them. So I felt that was a differentiator and we needed a differentiator in a such a crowded market. So I go back to that for my kid too. Like, okay 10 years from now I have I'm with my daughter and my daughter wants to be with me so much and we we play all the time and we are a 10 million AR company I would rather live in that reality than live in a reality where we're a hundred million dollar company and my daughter doesn't want to play with me it's not interested in me and we're not that close so that's the the decision I made. Maybe it's right. Maybe it's wrong. I have no idea. I go extremely hard.

I have a lot of ambition that's not going to stop me or anything. But like setting up those priorities and figuring out for yourself, like what works for me, I think it was important. So yeah, more than ever, like she's the most important thing for me. And I prioritize her a lot. how do you balance that with your ambition right like i also have this in a smaller way i don't have kids but uh yeah i have a lot of ambitious things and i have a family that i should spend time with and when i spend time with them i'm not present i'm looking at my phone yeah i'm like saying like oh i should be reviewing astra you know gpt6 like it's just like you know agi no it's not right like it's like but i really feel like i should be working and you know and like yeah at the same time, like, I feel guilty, and so, like, I'm happy in both areas. Totally, and that's how I feel. I feel like I'm a terrible dad, I'm a terrible husband, I'm a terrible CEO, and, like, you're failing in every single area of your life. I feel that way

exactly like that, but every day, I try to be a little bit more present, you know, and it's a, especially when you're working from home, like, man, you leave your office, and then now you're a dad and then now you're not anymore oh now you've got to be a husband and pay attention to your wife like and your your mind is like oh man how i do this enterprise like i gotta close this deal i feel like i suck all the time too we're on the same boat yeah yeah well never easy and i think like naming the thing for your category is important and people don't really think about category creation enough i think uh this is something we really do something like that netlify tried to do and failed with Jamstack. And I think that more people should really actually try to do a category creation because when they're successful, then they own the whole thing, which we did for this. Recent, you know, it is the next generation SendGrid, let's call it, right? Or like AI native SendGrid, something like that.

I think that, you know, and this is me being probably too honest, that is not category creation yet. Yeah. But it is category dominance. You own your category, that's good. But it is not clear enough yet what is different. So my classic example is another guest that we've had, Dormesh Shah from HubSpot. Anyone who wants to think about category creation, go and read up about inbound marketing versus outbound marketing and how they created the category of inbound. The ideal phrase of category creation is once you've heard the category, the two words that your company represents, you don't want anything else. there's no long essay there's no long manifesto there's no talk it's just two words maybe some examples that's it that your audience can take it from there it's very hard man i've seen friends trying to do that like svix uh from uh it's a web hooks as a service yeah yeah they took my name yeah man it's just so freaking hard right like i think they're doing okay right i don't know if

They're still around. There was a webhooks company, and then there was an anti-webhooks company. You were like, we don't do webhooks. We just help you publish JSON file. Oh, wow. It kind of reverses the push versus pull, the data flow. Yeah, this is, again, COVID-era DevTools. That was the most exciting thing there at the time. Cool. Well, always a pleasure to catch up, and congrats on all the success. I still, you know, am very curious how much you have dominated your category. I think you just chose it very well and you executed it very well and you're managing your family as well. So lots of respect. Thanks so much. Yeah, it was great chatting with you guys. Thanks a lot.