A Sparkling New Era: Everything We Launched at the Databricks AI Summit 2025

It's here. And it's big. A new version of Soda just dropped and you don't want to miss it. So get ready to get a sip of great data thanks to the newly found collaboration between Soda and Databricks.

This new version supports full integration with Databricks environments and all features have been tried and tested by Databricks advocates themselves. This thing can process 1.3 billion records in a little over a minute (64 seconds, to be exact). We showed it to the world last week at Databricks Data & AI Summit, June 9–12, 2025.

This launch is very dear to our hearts. As a data company, we know the struggle of data teams. We see the issues and needs they face daily; but we also envision solutions. To every "common pain point" in the data industry, we see an unrealized promise. And we think it's about time that the promise actually delivers. The new version of our software focuses on the very foundations that got us inspired and started in the first place:

What's New?

Here’s everything we shipped and announced last week:

• Acquisition news.
• The fastest data observability platform. To easily scale your data quality.
• The first collaborative data contracts. Enabling seamless collaboration between business and engineering.
• Transparent pricing. Because we’re all tired of the “book a call” button.
• And finally, the Soda Swag Store. Because we all love some fun merch.

Soda has Acquired ML Monitoring Startup NannyML

Together, we’re building the most intelligent, context-aware data quality platform on the market. One that helps you prevent issues before they become business problems, detect anomalies that actually matter, and trace root causes across the entire stack, from data ingestion to automated decision-making.

This move brings together two teams with a shared goal: helping data and AI teams ship reliable, production-grade systems they can trust, whether those systems power dashboards, models, or autonomous agents.

→ Learn more about the acquisition at: https://launch.soda.io/blog/soda-acquires-nannyml

The Fastest & Most Accurate Data Observability

Data observability focuses on starting right to, then, improving upstream. What went wrong in the past? What didn't we catch with our current checks? Has data changed without anyone noticing? Observability is all about making intelligent decisions based on what can be learned from the past.

Metric Observability Dashboard

Powerful algorithms are great, but equally important is how we use them to leverage results. The new Metric Monitors Dashboard brings data observability to life with a clean, intuitive interface designed to make complex trends immediately understandable. This visual tool acts as a health panel at a glance, built to help teams quickly identify and understand anomalies in key data metrics, and allowing them to act on the issues and shift-left.

Out of the box, Soda captures essential metrics, charts their trajectory, and flags unusual shifts the minute they happen. The result is instant, shared visibility: data engineers get a precise signal they can trust, while business teams see clear indicators. All without a line of code.

This feature was designed to solve a common challenge we kept seeing over and over again: anomalies that go unnoticed or are buried in noise. Now, from the first scan, the dashboard establishes a living baseline for “normal” and surfaces deviations in a clean, card‑based layout. Historical backtesting reveals whether today's spike is truly new or part of a trend, and users have access to opt-in alerts that notify data owners when something unusual happens. This way, deviations can be tackled before they make it into production, and data engineers can learn from historical data to prevent future issues.

At the end of the day, this isn't just a panel to view the system's health or to detect what went wrong. It's a tool for maintaining what's right and, then, shift left.

Powerful Anomaly Detection Algorithm

Anomaly detection takes a whole new meaning with the latest Soda version. We've developed a brand-new set of algorithms entirely in-house without relying on third-party black-box solutions to avoid rigid modeling assumptions and lack of interpretability. Designed to minimize false positives and missed detections, it shows a 70% improvement in detecting anomalous data quality metrics compared to Facebook Prophet across hundreds of diverse, internally curated datasets containing known data quality issues. This level of performance is critical in production environments, where missed anomalies and false alarms can undermine trust and introduce operational inefficiencies.

The key feature here is complete control and full transparency in the modeling process. This allows teams to explain the model's behavior and deliver improvements over time, adapting dynamically to your data and enabling precise, contextual anomaly detection. The algorithm features high accuracy while leveraging historical data out of the box.

This model starts analyzing data as soon as it's loaded and it can automatically learn from old and new patterns as new information makes it into the source. For those who prefer a more hands-on approach, it also supports human-in-the-loop feedback for continuous improvement and refinement.

→ Learn more about the new metrics observability at: https://launch.soda.io/blog/data-observability

The World's First Collaborative Data Contracts

The Collaborative Data Contracts editor will allow to toggle between code and no-code mode when creating and editing contracts, which is perfect for both technical and non technical users looking to update, suggest or implement changes.

A very common issue that data engineering teams face is spending too much time in meetings figuring things out with business teams. The other side of the coin? Non-technical people being unable to easily make suggestions or fully understand the data they're consuming.

A unified editor with multiple solutions for multiple technical levels is the perfect tool to expedite exchanges. You don't know code but need your data to follow new guidelines? Just suggest a new check with drop-down menus and a few clicks. No time to have a meeting with a business team to figure out their data needs? Ask for their vision and then tweak the code in your existing workflow so that it matches expectations. All contracts—whether created via code or UI—are interoperable and synchronized, ensuring consistency and shared understanding.

The Soda Cloud editor is here to push forward the main vision of this launch: bridging the engineering-business gap.

→ Learn more about Collaborative Data Contracts at:  https://launch.soda.io/blog/collaborative-data-contracts

A New Contract Language for a Unified Experience

We decided to revamp our configuration language and introduce a modern contract definition language, called Soda Contract Language, designed to ensure clear accountability and alignment between data producers and consumers, integrating seamlessly into the no-code interface to support non-technical users. Its structured and predictable foundation allows for readability and transparency, as well as simpler debugging.

If you're a data engineer, you'll be happy to hear that it's fully YAML-compliant and validated against JSON Schema, which improves flexibility and usability in all environments.

Transparent Pricing

We believe data quality should be accessible, not gated behind demos, opaque pricing, or sales conversations. Whether you’re a solo engineer managing three pipelines or a platform team running AI-driven workflows, you should know exactly what you’re getting, what it costs, and how fast you can start. That's why we decided to create a free plan and make our prices transparent.

You operate under pressure. You’re moving fast, owning more of the stack, and being asked to build systems that are observable, automated, and AI-ready. But the buying experience for data tools hasn’t evolved. You still have to guess what things cost.

So we’re cutting through that.

Our new pricing is designed to be simple, flexible, and fair:

• Start free: no credit card required
• Pay for what you use: per dataset monitored
• No seat-based pricing: unlimited users
• Clear upgrade path: scale when your needs grow

We’re aligning our business model with how modern teams actually build and scale data products.

→ Learn more about our Transparent Pricing at: https://launch.soda.io/blog/transparent-pricing

The Soda Swag Store

On the last day of the launch week we decided to launch something completely different. Something physical. We introduced The Soda Swag Store. The place where you can buy actual Soda sodas.

This merch drop isn’t just for fun. It’s for the community that’s been behind us from the start.

Both Soda and NannyML were founded on open source.

• Soda Core is the foundation for many teams’ data quality pipelines.
• NannyML’s OSS library has set the standard for post-deployment model monitoring.

We’ve kept these projects open, maintained, and community-first and we’re doubling down.

This merch store is our way of showing love to the contributors, users, and advocates who help us build better tools every day.

And we’re not done.

In our next launch, we’ll be releasing:

• Soda Core v4 - the most powerful version yet, with major improvements in flexibility, extensibility, and performance
• NannyML v1 - distilling 4 years of learnings into the most efficient version of nannyML OSS yet, minimal dependencies and blazing fast

→ Check out our new Sodas and Swag at: https://shop.soda.io/

In Short

The new version of Soda is here to build a bridge between data teams and business teams. Our vision is to give an all-in-one tool that can be completely incorporated within everyone's existing workflows. That's why we decided to support full integration with the platform of choice for the majority of data teams: Databricks.

We want to ensure that Soda meets users where they already are, seamlessly fitting into the tools they trust and use every day, so if you're not a Databricks user, do not fret. Starting June 30th, the new version of Soda will support full integration with all major data sources.

Through this launch, we're introducing key updates in our product:

• NannyML's Acquisition: We joined forces to manage the world's automated decisions.
• Metric Monitors Dashboard: A nice, visual dashboard to allow teams to analyze the performance of their data over time.
• Powerful Anomaly Detection Algorithm: Fully home-brewed, highly accurate, easy to adjust and incredibly transparent.
• Collaborative Data Contracts: An as-code and no-code solution for all, to ensure transparent collaboration. Because you can’t prevent what you don’t agree on.
• New, Simpler Contract Language: Clear, structured and built for easy alignment between teams and smooth integration into any workflow.
• Transparent Pricing: Because we are all tried of the book a call button.
• The Soda Swag Store: The place where you can buy the official drink for data engineers and more.

We hope you enjoy this new flavor of Soda. These updates aim to make everyone's job easier through a collaborative platform that brings teams together, simpler structures that allow for integration into any existing tool and workflow, and a clear separation between testing and observability, which clearly defines ownership and strategies without blocking workflows. This way, teams can move faster, collaborate better and trust their data more than ever.

Did you hear we're giving away a $1000+ mechanical keybaord?

Winning is easy; sign up for the new Soda Cloud to enter the raffle. 5x your chances to win the keyboard by completing our onboarding. Simply create your account here, our onboarding takes just a few minutes.