Streaming data powers the most important decisions in modern businesses. But building streaming pipelines still too often requires specialized expertise, manual SQL authoring, and time-consuming trial and error.

That changes today.

We’re excited to introduce DeltaStream Agent, a new capability that lets you use natural language to build fresh context and real-time data pipelines in DeltaStream. Instead of hand-writing DeltaStream SQL or manually stitching together streams, changelogs, joins, aggregations, and materialized views, you simply describe what you want in plain English. DeltaStream Agent figures out the rest.

This is a big step forward in making real-time data engineering dramatically more accessible, faster, and more intuitive.

From idea to live pipeline, just by describing it

Here’s the promise of DeltaStream Agent:

You tell it what you want to build.

It understands your data, inspects schemas, creates a plan, generates the required streaming logic, and helps deploy a working pipeline, all while keeping you in control with confirmations at key steps.

Video thumbnail
Play video

In the demo shown in the video, the agent was given a straightforward request:

Create a pipeline that enriches page views with user data, writes that enriched data to a new topic, and then builds a materialized view showing page views per zip code per minute.

From that one prompt, DeltaStream Agent:

  • Identified the existing Kafka topics
  • Inspected the data to understand schema and structure
  • Created a stream for page views
  • Created a change log for users
  • Defined an enrichment query joining the two datasets
  • Wrote the enriched result to a new topic
  • Built another streaming query on top of that topic
  • Applied a tumbling window for one-minute aggregation
  • Materialized the result into a view
  • Summarized the completed end-to-end pipeline

Then the resulting materialized view was queried to validate that real data was flowing through.

That is a complete streaming pipeline, built from natural language.

Why this matters

DeltaStream Agent is not just a convenience feature. It changes how teams can work with streaming data.

For years, creating real-time pipelines has required deep familiarity with streaming systems, query semantics, schema design, and platform-specific syntax. That expertise is valuable, but it has also created a bottleneck. The number of people who need fresh context keeps growing. The number of people comfortable building it from scratch does not.

DeltaStream Agent closes that gap.

It allows engineers, platform teams, data teams, and eventually a much broader range of builders to move from intent to implementation far more quickly.

The biggest benefits of DeltaStream Agent

1. Build pipelines in plain English

You no longer need to start with syntax. You can start with the outcome you want.

Instead of thinking, “What DSQL do I need to write?” you can think, “What data product do I want to create?”

That shift is powerful. It lowers the barrier to entry for new users and accelerates experienced users who would rather spend time on business logic than on repetitive plumbing.

2. Turn raw data into fresh context faster

Modern applications, analytics, and AI systems all depend on context that is accurate and current.

DeltaStream Agent helps you create that fresh context quickly by automatically translating a request into real-time streaming logic. Joining datasets, enriching events, aggregating over time windows, and materializing results becomes much simpler.

That means less time spent wiring systems together manually and more time delivering usable, high-value data products.

3. Let the system understand your data for you

One of the hardest parts of building pipelines is not writing the final query. It is understanding the underlying data well enough to design the right pipeline.

DeltaStream Agent helps by discovering existing topics, inspecting schemas, and using that information to build an execution plan. It does not just generate code blindly. It reasons about the available data and determines the right DeltaStream objects and transformations to create.

4. Go from prompt to working deployment

This is not just a text-generation demo.

DeltaStream Agent takes the process all the way through to execution. It can create the streaming resources, define the logic, and produce a working pipeline that you can validate immediately.

That end-to-end flow matters. It turns natural language from a nice interface into a real productivity multiplier.

5. Keep users in control

Automation should accelerate teams, not surprise them.

At each step where DeltaStream Agent needs to make a change, there is a confirmation. That gives users visibility into what is happening and control over what gets deployed.

You get the speed of an agent with the confidence of guided execution.

What DeltaStream Agent unlocks

DeltaStream Agent makes it much easier to create the kinds of real-time pipelines teams need every day, including:

  • Event enrichment
  • Real-time joins between streams and reference data
  • Derived topics for downstream applications
  • Windowed aggregations
  • Materialized views for operational analytics
  • Fresh context pipelines for AI agents and applications

This last point is especially important.

As AI agents become more central to how teams work, the quality of their output increasingly depends on the freshness of the context they receive. DeltaStream Agent helps users build the pipelines that continuously create and update that context, without requiring them to become streaming experts first.

In other words, DeltaStream Agent does not just simplify pipeline development. It helps accelerate the broader move toward real-time, context-aware software and AI systems.

A better way to work with streaming data

The traditional approach to streaming development often looks like this:

Understand the source topics. Inspect schemas. Decide which DeltaStream objects to create. Write DDL. Write queries. Check semantics. Deploy. Debug. Iterate.

DeltaStream Agent compresses that process dramatically.

You describe the outcome. The agent handles the discovery, planning, generation, and execution support needed to get there.

That means:

  • Faster prototyping
  • Faster onboarding
  • Less manual effort
  • Fewer platform bottlenecks
  • Broader access to real-time pipeline creation
  • More time focused on business outcomes

This is just the beginning

DeltaStream Agent represents a new interface to DeltaStream: one where users interact with the platform at the level of intent, not just implementation details.

That is a major leap forward for usability, and it opens the door to a future where building real-time systems feels far more natural, iterative, and accessible.

If you have ever wanted the power of streaming data without the friction of building everything by hand, DeltaStream Agent is for you.

Try DeltaStream Agent now

If you want to see how quickly you can turn a plain-English request into a working real-time pipeline, now is the time to try DeltaStream Agent.

Describe the pipeline you want. Let DeltaStream Agent discover your data, generate the streaming logic, and help deploy it for you.

Whether you are enriching events, creating live aggregations, building fresh context for AI, or standing up entirely new streaming workflows, DeltaStream Agent can help you get there much faster.

Call to action

Stop hand-coding every pipeline. Start building real-time data products at the speed of thought.

Request a demo, get early access, or try DeltaStream Agent today to see how fast natural language can become fresh context in production.

Share what you want to build, and let DeltaStream Agent do the heavy lifting.