An update to Gretel’s license to support continuous community growth and innovation
Gretel is changing the license for some of its open source components that also are a part of Gretel's SaaS Platform. We are migrating from Apache 2.0 to the Gretel Source Available License (SAL). This new license allows you to freely install and incorporate Gretel software packages into your own business, research, academic, or personal solutions. You may also modify and create derivatives, but may not redistribute to third parties or build a software-as-a-service offering with the packages.
Additionally, we will be adding a Contributor License Agreement (CLA) that must be accepted if contributions are made to the affected software packages. Similar to licensing agreements that other products like TensorFlow and Pytorch use, Gretel’s CLA asserts that contributors have the right to make a contribution, that contributors may maintain any copyright or patent protections for their contributions, and that Gretel and any other licensees and users of Gretel products may utilize the contributed code.
The two software packages impacted by this change are:
- Gretel Synthetics, starting with version 0.19.0. This package contains Gretel's core code for creating synthetic models and generating synthetic data.
- Gretel Trainer, starting with version 0.3.0. This package contains some of our preview features and advanced client interfaces that integrate with Gretel's SaaS Platform.
For more information, please see our FAQs on the SAL and CLA.
Why are we doing this?
We believe this is a necessary step to ensure that Gretel can continue to invest heavily in cutting-edge synthetic data capabilities that are freely distributed for the community, while sustaining a healthy business that funds these endeavors for the foreseeable future.
This approach is common amongst other data-centric platforms such as Confluent, Elastic, and Redis. These businesses have launched their own "as-a-Service" offerings and have also been simultaneously working with (or battling with) major cloud providers that have varying approaches regarding the use of open-source software. Fortunately, these data-centric companies have been very transparent in their dealings with these cloud providers, and as Gretel begins onboarding our own large strategic partners and enterprise customers, we are taking a similar stance of proactivity instead of reactivity.
Of course, we could have chosen to never release any of our core synthetic capabilities as open source from the very beginning and taken a more proprietary stance. But what good would that be? My co-founders and I have already played this game, coming from the cybersecurity world where every algorithm is "next generation" and "bleeding edge," and the products operate in a black box – obfuscating basic functionality from the users they are intended to help. No thanks.
The synthetic data market is emerging, and we have an opportunity to keep it honest, collaborative, and innovative while building our business at the same time. Publishing our core capabilities for the community to use encourages transparency, fosters open research and collaboration, and, most importantly, earns trust with our users, whom we view more as peer developers than customers.
What do you need to do?
If you're already a Gretel SaaS Platform user, we love you! Keep on keeping on. If you are using Gretel Synthetics or Gretel Trainer, we also love you! For the most part, the users of these components should continue to use them exactly how they have been – incorporating Gretel packages into their solutions, enabling the use of synthetic data for their research or business.
If you are considering using Gretel, in any capacity, to offer synthetic data capabilities to your own customers or users, does this mean you are out of luck? No! Just contact us and we'd love to work or partner with you on your use case. This licensing update just means you can't outright take our software packages and build your own business without speaking with us first.
Now get synthesizing!