We're very excited to announce Convex, a new company empowering software developers to build sophisticated, highly dynamic serverless applications with ease.
Earlier this year, Convex raised a $3.5M seed round led by Neo, with additional funding from Fathom, SV Angel, General Catalyst, and #ANGELS. We're also proud to have the support of a network of more than 80 individual angels, including partners from Andreessen Horowitz, Sequoia Capital, and Kleiner Perkins, as well as nearly a dozen founders of technology unicorns.
Recently, we've been busy hiring our initial team, demoing the alpha platform to engineers in our network, and getting ready for a private beta launch later this year.
How we got here
Prior to Convex, for eight years Convex's founders worked closely together at Dropbox, leading many of the company's major infrastructure projects and teams. That experience led to a lot of lessons on scaling a large technology company—lessons on architectural choices, processes, talent, and cultural values that are essential to success.
After leaving Dropbox, we were eager to use the public cloud to quickly prototype a bunch of apps to validate our early business ideas. But when we started looking in depth at those clouds' management consoles, we were, to be blunt, stunned.
The public cloud is still just queues, VPCs, and databases. There are routing tables to misconfigure and availability zone failure domain methodologies to reason about. We weren't interested in making decisions about which distro and kernel to use, nor did we want to interrupt our flow to configure lots of very granular IAM policies.
We just wanted to build products. The language of the cloud is really far away from the customer workflow-level language we wanted to live in.
Upon reflection, to us, it feels like there's been a bottom-up process in which these huge companies have exported their org charts. "We have a storage team, now you have a storage team. We have a network team, and now so do you."
But we still have to glue all this stuff together, and most ways to do so are the wrong way. To discover the right ways, it is still necessary to develop an incredible level of expertise—expertise that has nothing to do with your team's product.
So we paused our coding, talked to a lot of other startups, and found most of them shared very similar frustrations. Many had used solutions like Firebase early in their project and considered the day they outgrew it a very sad one. "We're back to doing DevOps and having on-calls... this stuff just has nothing to do with improving our product. And there's nothing different about how we do it than any other company."
The coming serverless revolution
In 2021 the serverless movement is about to fundamentally change the way teams build software on the internet. It's taking away difficult, non-differentiated work. There is some progress now, but much is still needed.
The existing serverless tools and platforms today provide solutions for fairly simple sites and teams. For projects which are mostly static content and web only, it's now possible to create and maintain these without worrying about the details of how servicing the content at scale actually works. But as soon as content gets user-driven and dynamic, you're quickly back to gluing together thinly-papered-over same-old systems.
Dynamic sites really deserve the same hands-free relationship to scale, performance, cost efficiency, and reliability that static ones currently enjoy. But making this seamless requires a more radical, comprehensive rethinking of how backends should work moving forward.
So at Convex, we're designing a new set of abstractions that are based on a top-down examination of how web and mobile engineers work today—not a bottoms-up export of an org chart from 2003.
We've built natural and ergonomic SDKs for storing and querying data, performing efficient computation, subscription, and caching. Soon, we'll introduce ways to trigger and manage asynchronous computation (timers, queues, pipelines, etc.) and provide frameworks to gracefully evolve your system (gradual schema enforcement, 100% safe data migrations).
Our goal is to end up in a world where an engineer, even on a large team and codebase, can make some changes on their laptop, verify everything works locally, press the big green "deploy" button, close their laptop, and go out to dinner with complete confidence their changes will work identically at scale.
The first puzzle piece
Later this year, we'll release the first closed beta of our platform. Our first release will contain the core capabilities that empower web developers to develop highly dynamic applications with shared user data faster than ever before.
Stay tuned for more details in the next few weeks. And if you're interested in getting an early invite, sign up for our beta waitlist.
And if you'd like to help us build this future, we're hiring! We're looking for people who are passionate about improving the development experience for frontend engineers and who have strong opinions on what kinds of tools and workflows will remove barriers to getting things done.