← All posts

Announcing $2.6M seed investment

The round is led by 645 Ventures with participation from Y Combinator and several other firms and angel investors.

Header image
By Jai & Sanket on 
Share on Twitter Share on LinkedIn Share on Facebook

We’re glad to announce that we have raised a seed investment of $2.6M, led by 645 Ventures with participation from Y Combinator, FundersClub, Pioneer Fund, Liquid2 Ventures and several other angel investors.

In December 2018, we set out to build a product that helps developers automate all the objective parts of code review, using the power of static code analysis. Static analysis as a technology has been in existence for almost as long as programming languages. Still, the adoption has been abysmal in general so far. We strongly believe static analysis can add value to developers in saving time and improving the quality of their code when it is easier to set up and use, carefully integrated with code review workflows, and has a very low rate of false positives.

Developers waste over 17 hours on an average in a 41-hour workweek due to technical debt, bad code, and delayed code reviews. According to recent research, the global GDP loss from developer time wasted on bad code annually is pegged at $85B. At DeepSource, we’re solving for this by building an automated code review platform that helps developers detect issues in their code and automatically fix most of them in a couple of clicks with Autofix. We will be using this capital to invest primarily in product and engineering, developer relations to expand our reach amongst developers, and building a community of developers that care about writing good code.

First release

We released the initial version of DeepSource that supported only Python with GitHub integration in Feb 2019 on Product Hunt. With support for only open-source repositories, DeepSource was completely free to use. For this first release, we had these three goals in mind that aim at solving the most critical problems with static analysis tools in general:

  • Simple configuration
  • Integration with existing code review workflows
  • Less than 5% of false positives

This launch got us our initial set of developers. We also received requests to support providers like GitLab and Bitbucket, and also analyze Go, Ruby, and JavaScript.

Initial adoption by open-source projects

In May 2019, we listed events in San Francisco where Python developers hung out and we attended almost all of them. Meeting developers and onboarding them in person was pivotal to how the product has evolved so far, and it allowed us to put our ears very close to the ground. Some popular open-source projects started using DeepSource, and the maintainers provided us with regular feedback. The Ludwig team at Uber even wrote about us in the Uber Engineering release blog.

Uber Engineering blog

Pre-seed fundraise

With the initial product in place, which was used by open-source teams at Uber, Slack, NASA, amongst the others, we raised a pre-seed round of $140K from angel investors. Bradley Buda (Founder, Census) wrote the first cheque along with John Kinsella (Ex VP Engineering, Qualys), Badri Rajasekar (Founder, Jamm), Bhavya Sahni (Marketing, Wingify), Smruti Parida (Ex CTO, Nestaway), Sanjay Suri (CTO, Nykaa), Seth Bindernagel (VP Marketing, SoloLearn) and Soso Sazesh (Founder, Growth Pilots). In the coming months, we released support for analyzing Go, Dockerfiles and Terraform and added integration with GitLab. Projects like DGraph, Gauge (ThoughtWorks), among others, started using the Go analyzer as part of their day-to-day code review workflows.

Y Combinator

In November 2019, we were accepted to Y Combinator’s Winter 2020 batch. YC has been a huge turning point for DeepSource. Around this time, we had integrated the billing module with Stripe and started supporting private repositories. We saw many teams integrating DeepSource into their team’s private repositories. During YC, we learned from our partners and alumni how to position the product to developers, and we also learned product improvements to increase developer adoption organically.

We discovered that one of the primary challenges our users faced was fixing all the occurrences of the issues DeepSource raised manually. This was a tedious process in projects with a large number of files. This accelerated us towards building Autofix, which we had envisioned as the next step in pushing the limits of what can be done with static analysis. We released Autofix in February 2020, which gives us the ability to fix erroneous code in a couple of clicks automatically.

DeepSource Autofix

Seed Investment

645 Ventures lead our seed investment of $2.6M, with participation from Y Combinator, FundersClub, Pioneer Fund, Liquid2 Ventures, Christopher Golda (Rogue Capital), Timothy Chen (Essence fund), Ivan Kirigin (Tango VC), Ed Roman (Hack VC), Jakub Jurovych (Founder, Deepnote), Mike Viscuso (Co-founder, Carbon Black), Venture Souq, Tokyo Black, Bradley Buda (Founder, Census), John Kinsella (Ex VP Engineering, Qualys), Soso Sazesh (Founder, Growth Pilots). This new funding will help us continue to improve the product, add support for more programming languages, integrate with source code hosting tools, and improve coverage of Autofix issues.

As of today, developers have fixed over 3.7 million issues reported by DeepSource. Since the launch of Autofix in February 2020, almost 10% of these issues were automatically fixed. The adoption of Autofix is growing rapidly, and we are working to automate fixing most of the issues we detect in the near future.

Resolved issues stats

Towards automated code reviews

Code reviews are completely manual today, even in engineering teams at large tech companies who use static analysis considerably. The typical workflow includes several steps that have to be done manually by the developers in the team — including code reviews, linting, security audits, and issue remediation. We believe that it doesn’t have to be like this — and that computers should be able to help us write good code.

DeepSource’s mission is to automate all the objective parts of code reviews and help developers write good code. The initial release of Autofix has been a small step in this direction. We have seen countless engineering teams start using automated fixing with it as a critical part of their code review process. We are working on many more exciting things that bring more automation and intelligence into static analysis and code reviews.

We want to thank our team and our users for helping us come this far. Keep an eye on us as we keep building tools that help you ship good code.

Stay safe!

About DeepSource
DeepSource helps you automatically find and fix issues in your code during code reviews, such as bug risks, anti-patterns, performance issues, and security flaws. It takes less than 5 minutes to set up with your Bitbucket, GitHub, or GitLab account. It can analyze Python, Go, and Ruby code. JavaScript is coming soon.
Sign up for DeepSource