Software development is an interesting task for many reasons, but among the worst and most wasteful parts is waiting for containerized builds to run and test their applications.
Developer Kyle Galbraith found himself twiddling his thumbs and waiting for software builds too often, so he was inspired to find a way to make them much faster.
That led him to co-found Depot, a rapidly growing build acceleration platform, which today announced its $4.1 million seed funding round led by Felicis, with participation from Y Combinator, Aviso Ventures, Tokyo Black, and several angel investors.
“Today, Depot is a build acceleration platform that makes container image builds and GitHub action workflows up to 40 times faster,” said Kyle Galbraith, CEO and co-founder of Depot, in a video call interview with VentureBeat earlier this week.
This capital infusion is set to propel Depot’s expansion and further enhance its platform, which is already making waves in the software development community.
Depot’s ultimate vision is to make all builds—whether local or in continuous integration—near instant, with comprehensive ecosystem integration. “The ultimate goal here is to make builds, all builds, whether they’re local or in CI, or whether they’re Docker or GitHub action runners or something else, to make all of those near instant with comprehensive ecosystem integration,” Galbraith said.
The company is using the same Linux VMs as others but is optimizing them with its own tools and services to deliver generically faster build processes across the board.
Depot’s origin story
Founded in 2022 by Galbraith and his co-founder Jacob Gillespie, Depot was born out of their experience as platform and DevOps engineers.
“We started building Depot back in 2022 to effectively make our lives easier as platform engineers,” Galbraith explained.
Depot was created to address the inefficiencies and slow performance of container image builds within GitHub Actions, a common pain point in software development.
By putting the build kit on cloud virtual machines (VMs) and persisting layer cache automatically to real SSDs, Depot was able to orchestrate across builds, initially making them five times faster initially.
Rapid scale-up
Since its inception, Depot has rapidly scaled its operations. After launching its beta in 2022, the company attracted 10 customers in its first week and gained another 10 within just two to three weeks of its full product launch in July 2022. Depot then joined Y Combinator’s Winter 2023 batch, further accelerating its growth.
Depot’s platform is now helping over 1,800 organizations, processing 1.3 million builds monthly.
“We see 1.3 million builds a month today, so we’re sitting on a wealth of information already to help inform [an AI] model,” Galbraith noted.
Notable clients include PostHog, Wistia, and SendGrid, which have adopted Depot’s solution to streamline their software development processes.
Depot’s platform offers developers speed and cost-savings
The platform leverages optimized Linux VMs with proprietary tools and services, supporting native Intel and ARM builds.
Additionally, Depot is popular with AI companies that build and package large language models (LLMs), as well as with traditional web app developers and application service providers.
Galbraith emphasized the ease of use and performance benefits that Depot offers. “For container image build acceleration, you create a Depot account, and you literally swap Docker build for Depot build, and you’re done.”
The platform optimizes context transfer to send only changed files between builds, and it has no discovered upper limit on application size, supporting the building of several terabyte LLM model files.
“We have yet to discover that limit today, especially around LLMs, because some of the things that people are building on Depot, they actually can’t build locally,” Galbraith added.
Dev-centric approach
Depot stands out not just for its performance but also for its customer-centric approach. “It’s really fun talking to Depot customers, because they’re like, ‘Oh, I have this one small, really annoying thing with this provider,’ and we’re like, ‘You should 100% tell us, because we’ll probably go fix it, because we hate paper cuts,’” Galbraith shared.
The company’s billing and security practices are also tailored to address common pain points in the industry. “We do a lot of things that we have not liked with other providers. For example, we bill by the minute, but we count by the second,” he said.
Additionally, Depot ensures security by isolating each build in a dedicated VM. “Unlike other providers, we draw the boundary at the VM. So when you run a build on Depot, whether it’s a GitHub action job or a Docker image build, we give you the entire VM to that build so there are no other customers next door,” Galbraith explained.
What Depot will do with seed funding
Depot’s seed funding will help the company expand its offerings, focusing on new build inputs that can further accelerate development processes.
The company is currently expanding support to include macOS and Windows environments, which will broaden its appeal to a wider range of developers.
This expansion is a critical step in Depot’s mission to optimize software builds for developers worldwide. The company also plans to partner with other services, such as Fly.io, and develop AI-powered build optimization suggestions using a fine-tuned model.
Jake Storm, a Partner at Felicis, highlighted the potential impact of Depot’s technology on the broader software development industry. “Depot is scaling the build acceleration platform that will revolutionize developer productivity in all environments,” Storm noted. The platform’s ability to streamline both local and continuous integration builds positions it as a transformative tool for developers seeking to enhance their workflows.
Depot’s rapid growth and innovative approach to solving build inefficiencies have made it a standout in the tech industry. As the company continues to expand its platform and reach, it is quickly becoming a critical piece of infrastructure for developers aiming to speed up their software development cycles.
Source link
