Today, we’re excited to announce Firebolt Core. Firebolt Core is a forever free, self-hosted edition of Firebolt’s distributed query engine. It provides high-performance data warehousing capabilities that can be deployed anywhere you’d like to deploy it. Core is designed for engineers who need low-latency, high-concurrency analytics. You can use it for one-off ad-hoc analysis, to power 24/7 production workloads, or anything in between. The only thing it cannot be used for is to build a hosted SaaS solution which competes with Firebolt’s managed service.

Firebolt Core contains most of the features present in Firebolt’s managed service. Most importantly, the query engine in Firebolt Core is exactly the same as in the managed version of Firebolt. Without any handicaps or limitations. And it is free. Forever.

What Is Core?

Firebolt Core is a production-grade, distributed query engine you can run locally on a single laptop, in the cloud, in a massive datacenter, or anywhere in between. It comes with no cost and no lock-in. It is incredibly easy to get started with Firebolt Core. Just type the following to start a local Core instance on your machine and drop into the CLI.:

bash <(curl -s https://get-core.firebolt.io/)

You can find Helm charts to run a multi-node setup in our GitHub repository.

What makes Firebolt Core particularly exciting to us is that it's not a stripped-down or limited edition of Firebolt. It's the same engine that powers our cloud platform: a fully distributed, vectorized query engine with aggressive indexing and modern SQL support. It's built to serve large analytical workloads and can handle the same scale whether you're running on a local dev machine or across a Kubernetes cluster. We believe that there is no real alternative in the market today. With Firebolt Core, you're not just getting a taste of the engine, you're getting the full experience.

What Can You Do With Core?

Firebolt Core excels at handling a wide variety of analytics workloads and it is built to handle production workloads: the query engine is the same as the one used by Firebolt’s production customers today.

Firebolt Core can power data and AI applications and take on highly-concurrent analytical workloads. With Firebolt’s optimizations and features like aggregating indexes and subresult reuse, you can achieve millisecond latency for your queries on Core. We use Firebolt Core for our Clickbench submission, where as of the time of publishing this blog, it holds the #1 spot. 

You can also power your ELT workloads at scale. Core integrates natively with Iceberg, and is easy to set up for scale-out processing. This empowers you to build a wide range of high-performance analytics applications.

Why Did We Build Core?

Why give away the same high-performance engine we've spent years building? Because this move actually makes business sense for us. We believe it will mutually benefit both the engineering community and Firebolt.

Data architectures are becoming more modular. More teams are opting for open table formats, decoupled storage and compute, and hybrid environments that span public cloud, private cloud, and on-prem. Firebolt Core fits into this vision cleanly. It supports Iceberg out of the box and plays nicely with object storage. It’s a flexible foundation that you can adapt to your stack, rather than the other way around.

This makes it easy to adopt Firebolt for modern software projects. No matter whether you start your journey on a laptop, a VM, or a local cluster, you can leverage Firebolt from day zero. For projects that later decide they’d rather not manage infrastructure or want seamless scale and support, Firebolt’s fully managed cloud offering will be there. We want you to start where you are, and grow with you at your own pace.

We’re living through a shift in how infrastructure is consumed and trusted. More than ever, developers want to try before they buy on their terms. With Firebolt Core, they can. It’s fast, free, and self-hosted. If you’ve been curious about what Firebolt can do, there’s never been a better time to try it.

Firebolt Core is built to respect the “try it before you buy it” paradigm and the shift towards increasingly technical and knowledgeable users who have no issues with self-managed solutions. Too often, powerful tools are locked behind paywalls, complicated procurement processes, or opaque platforms. We wanted to do something different: give you something you can use today, without needing to talk to sales or commit to a contract. We trust that once you experience the performance and simplicity of Firebolt, you'll understand why we built it.

We’re also hoping Firebolt Core becomes a platform for learning and innovation. Whether you're building an internal analytics stack, testing new data modeling approaches, or simply exploring what’s possible with Iceberg and SQL, Core is there for you. We’re excited to see what use cases and tools the community builds around it, and we plan to support and learn from those efforts.

Why Isn’t Core Open Source? 

We want to address the question that many of you have: why isn’t Firebolt Core released as an open-source project? This is something we’ve debated at length internally as well.

We want Firebolt to be a great business. With GenAI accelerating software development, defensibility is more difficult than ever. Our query engine is built on years of innovation in database performance by Firebolt’s engineering teams. GenAI can’t replicate that work yet, but open-sourcing it would make it easier to copy and commoditize. To protect that edge, we chose to keep Core closed-source.

We also want Core to be a full-blown version of our query engine that can power your production workloads. If we open-sourced Core, it would either have to be a limited version of the product, or we would hurt our ability to build a great business.

There’s another reason this decision makes sense: database lock-in is becoming less of a concern. Open table formats like Iceberg are democratizing data access, and we’re betting on that future. Core can directly read from Iceberg tables. With a single COPY TO command, you can export data to open file formats—and soon, directly to Iceberg. Firebolt aligns with Postgres for its SQL dialect, making it easy to move queries between systems.

This means that using Core doesn’t require an unreasonable leap of faith. We mean what we’re saying, but you don’t know us. But you’re also not betting your career on us, as moving away from Core is easy.

Engineers powering production workloads want a database that’s efficient, easy to use, gives them control, and doesn’t lock them in. We hope that for many of you, this will be Firebolt Core.

Conclusion

We’re releasing Core because we believe that giving away our query engine for free and achieving commercial success are not in conflict, but rather will be a virtuous cycle. Free doesn’t mean unsupported: it means inclusive. Firebolt Core is our way of giving to the community, helping to democratize high-performance analytics. We want to earn your trust, and look forward to building together.

If you want to chat with us about Core, join us on our Discord.