How MailKite Built an All-in-One Email Platform for SaaS Founders

After building my 5th or 6th SaaS, I kept having to recreate email. There was no email platform built for serial SaaS solo founders. Not only are you the developer, but you are also the marketing and support. Solo founders and small teams need three things from an email provider, an email API (send and receive - not just send), an agent API (MCP and agent skills), a UI for humans. I needed one simple platform for all three things I did daily building and marketing a SaaS. So I built MailKite for my own use and then figured I would like other developers to use it for their startups, to cover the missing surfaces that other transactional email platforms did not address. The main feature of the product is that it supports unlimited domains and team members for free, I always believe in having only real expenses that require payment, such as actually sending a large amount of email, not seats or domains. Now, with AI, founders are launching a product every few months and need a single platform that covers all.

Partnering with existing software with distribution channels is the largest targeted developer audience growth. Conversely, integrations with existing software platforms turned out to be much more difficult to gain users from when you do not have a distribution channel to feed the integration. Relying on a listing in a directory does not work - especially if the market is already saturated. Email providers are many; what separates MailKite from those does not guarantee traffic from a listing. The same would go for any other product that you list - do not spend time listing, spend time partnering. For example, building a plugin for WP to send email via SMTP with MailKite flopped. WordPress does not send traffic to your listing that has no distribution channel pre-existing. But partnering with an existing SMTP provider listed on WordPress that has a large user base brings in traffic. The same goes for NextJS or any other platform that sends email.

Honestly, the loudest complaint early on was just getting started. Setting up DNS, then sending and receiving that first email. People would message me asking what the code should actually look like, what the response comes back as, and they'd be stuck there before they ever saw the thing work. So I went after the docs. That's really been the fix. Since launch, I've made them a lot richer, quickstarts that get you from signup to a delivered email in a few minutes, copy-paste samples, boilerplates, SDKs for all the major languages, and I show the actual request and response payloads instead of making people guess. And I think this matters more than people realize. Developers judge an API by its docs. Nothing kills a good product faster than bad documentation. Agents read the same docs too, better docs, fewer mistakes, cleaner integrations, happy developers and agents.

Most people treat the email API and the human email platform as two separate markets. You buy an API for your transactional sends, and you buy a mailbox somewhere else for the humans. I think that split is wrong because email is ultimately human. Yes, you want automation, that's table stakes. But you also want a person to be able to look at what went out. You want visibility into the system. You want to step in when something matters, escalate a thread, reply to the customer who's about to churn. Right now, that's a silo. Your transactional email disappears into a log file nobody reads. And the same channel that sends your password resets is the channel you market through, and the channel where you can actually read the health of your SaaS. Depending only on code or only on AI means you're missing a huge part of how you reach the humans using your product.

The big push is partnerships, getting MailKite into the email software and platforms people already use. Here's the thinking. Most platforms only solve sending. That's the part everyone builds, and then they stop. They forget the rest of the surface, there's no real inbound API, so replies just vanish. There's no human interface, something that looks like an inbox but is actually built for a small team to manage email and handle support together. And now there's a third piece nobody's covering, which is the agent interface, giving an AI agent its own mailbox it can actually send and receive on. So rather than compete with those platforms head-on, I want to plug into them and provide the parts they're missing. They keep the sending, we complete the surface. Practically, that means integrations and provider plugins for the frameworks and tools developers already reach for, and then real co-marketing relationships with the platforms on top of that.

We have less than 20 paying customers, with just under 2K MRR. We have only been around for 3 months, so I feel it's a good trajectory. Marketing a SaaS is a long process - what matters is that I feel the existing customers are very happy with the platform and have started to explore many of the features that other platforms do not offer.

I've been building software for about twenty years, and I've started a few companies before this one, all pre-AI. I wasn't coming from inside the email industry, but I've been a customer of it over and over. Every product I've shipped needed email wired into it, and every time it was the same friction, pick a provider, fight the DNS, get sending working, and then discover that anything past sending, replies, a place for a human to actually look at the thread, was somebody else's problem. So the gap I'm building into is one I'd hit personally, many times, rather than one I found in a market report. What's different this time is scope. The earlier startups were sized to what I could build by hand. With AI, I can take on something considerably more ambitious, infrastructure, SDKs across every major language, deep documentation, work that would have needed a real team a few years ago.

This may not have been possible to learn before building the product - underestimating spam. I wanted MailKite to be very generous to the solo founder, allowing unlimited domains and team members for free and enough email quota for most startups. I was aware this would attract abuse - but not to the extent that it was abused. Spammers already had years of experience dodging safeguards, and a new email provider hasn't caught up yet. The first month was spent fighting spammers and building automated spam detection. Email providers are only as good as their IP reputation. A spammer can destroy an IP rep in a few hours. Luckily, with the state of machine learning and AI adding spam detection signals and flagging automated reviews and demoting spammers to lower IP rep servers, most of the IPs were protected.