For example, one of the world’s largest cloud computing and virtualization technology companies came to JFrog looking to implement their DevOps and DevSecOps cloud-first strategy. First looking to host JFrog platform services themselves in the cloud, they rapidly realized they would greatly lower their total cost of ownership due to the lower maintenance and infrastructure costs that would be safe by utilizing JFrog cloud services. With nearly 3,000 developers, this Company is working with JFrog to provide a complete end-to-end platform implementation, which they acquired through the Microsoft Azure marketplace through our strategic sales and partner teams. We believe our long-term investment in this top-down enterprise sales approach driven by our strategic team will continue to provide the North Star for JFrog’s top-shelf customers’ goals.
Finally, regarding AI and machine learning, we were proud to announce the availability of DevOps and DevSecOps industry-first support for machine learning model management in the JFrog platform in late Q3. As Katie Norton, senior research analyst for DevOps at IDC noted in her SwampUP report, quote, the ML model management capabilities expand the JFrog platform into an entirely new space and persona. A single system of record that can help automate ML models development, ongoing management, and security alongside all other components packaged into an application offers a compelling alternative, end quote. Just as we provided common ground for developers and IT engineers years ago when DevOps started, JFrog now has the opportunity to bring the world of AI and ML ops into the world of DevOps and DevSecOps.
Just as companies increasingly placed the demand on development teams to support security, they are now pressuring DevOps teams to deliver machine learning and AI capabilities in line with their applications. JFrog now provides a platform for binary management for developers, data scientists, machine learning engineers, and the machine themselves. As keynoted at SwampUP, JFrog is the first and only repository manager to proxy the most popular public AI and ML model repository, Hugging Face. We are also the first solution embedding security for ML models to scan models before they are used. Artifactory now natively supports AI models within the JFrog platform to manage and deliver ML capabilities alongside all other binaries within an organization.
The AI supply chain is a software supply chain, and we are proud to be addressing this demand on developers. With that, I will turn the call over to our CFO, Jacob Shulman, who will provide an in-depth recap of Q3 financial results, as well as update you on our guidance for Q4 and for fiscal year 2023. Jacob?
Jacob Shulman: Thank you, Shlomi, and good afternoon, everyone. During the third quarter, total revenues were $88.6 million, up 23% year-over-year. Our stronger than expected revenues in the quarter were driven by continued strength in our cloud business and expansion of large customers on the JFrog software supply chain platform. In the third quarter of 2023, our cloud business saw sequential expansion in customer usage, equalling revenues of $30.6 million, up 46% year-over-year. We’re happy to see the acceleration in our cloud business growth, driven by continued increase in cloud usage and better collaboration with public clouds to enable large customers on their marketplaces. We reiterate our baseline cloud growth rate of mid-40s during fiscal year 2023.
Self-managed revenues on-prem were $58 million, up 14% year-over-year during the quarter. We continue to see that our customers’ roadmaps and plans are focused on SaaS migration, and therefore, they do not expand their on-prem footprint at the same pace as in prior years. We continue to believe that our security solutions can be a potential catalyst to re-accelerate revenue growth and customer expansion within our self-hosted business. Net dollar retention for the four trailing quarters was 119%, a decline of one point sequentially. As predicted, NDR is stabilizing, and we continue to expect our net dollar retention rate for the year to be at these levels. Our gross retention continues to be 97%, with no change in overall customer return trends.
