Diffblue
Oxford University spinout that uses reinforcement learning to generate unit tests for Java code autonomously.
📍 United Kingdom 🇬🇧, Oxford
Product overview
Diffblue is an Oxford-based company that automates the writing of unit tests for Java code. Unlike code suggestion tools that provide snippets for developers to review, Diffblue Cover works fully autonomously: it analyzes existing code, generates complete unit tests using reinforcement learning, and guarantees that every test compiles, runs, and finds bugs. The tool runs on-premise, so source code never leaves the customer's environment. The company was spun out of the University of Oxford's AI research group in 2016 by Daniel Kroening and Peter Schrammel. It raised a $22 million Series A led by Goldman Sachs in 2017 and has since raised a total of around $47 million. Customers include Goldman Sachs, S&P Global, Citi, JP Morgan, and AWS. Diffblue Cover integrates into CI/CD pipelines and supports IntelliJ as an IDE plugin. The company offers a free Community Édition for individual developers, a Developer edition, a Teams édition (around $30,000 per year), and a custom-priced Enterprise edition. The tool focuses exclusively on Java but the company has signaled plans to expand to other languages. KEY FEATURES: - Fully autonomous unit test génération for Java using reinforcement learning - On-premise deployment with no code leaving the customer's environment - CI/CD pipeline intégration via GitHub Actions, GitLab, and Jenkins - IntelliJ plugin for real-time test génération during development - Free Community Édition alongside paid Developer, Teams, and Enterprise tiers