STREAM PROCESSING

Transform Data As It Flows

Apply transformations, aggregations, and enrichments to your data streams in real-time. Write in SQL, Python, or TypeScript—powered by Apache Flink.

Start Free Trial
Read Docs

Powerful Stream Processing

SQL Transforms

Write SQL to transform, filter, and aggregate streaming data with Flink SQL.

Python & TypeScript

Build transforms in Python (PyFlink) or orchestrate pipelines with TypeScript via API.

Windowed Aggregations

Compute rolling averages, counts, and other metrics over time windows.

Stream Joins

Join streams together or enrich with lookup tables.

Inline Inference

Run ML models and LLM calls on streaming events as a transform step. Sentiment analysis, classification, entity extraction, and scoring in real time.

VS Code Editor

Write and test transforms in a built-in VS Code editor with syntax highlighting.

GitHub Sync

Version control your transforms with GitHub integration.

Built-in Transforms

Apply these transformations to your data streams without writing code

Field Mapping

Rename, reorder, and restructure fields to match destination schemas.

Data Masking

Hash, redact, or tokenize sensitive fields for compliance.

Type Conversion

Convert data types automatically between source and destination.

Filtering

Filter out rows or columns you don't need in downstream systems.

Computed Columns

Add new fields based on expressions or lookups.

Flattening

Flatten nested JSON structures into tabular format.

Transform with SQL

Write familiar SQL to transform your streaming data. Filter rows, rename columns, compute aggregates, and join streams together.

  • Standard SQL syntax
  • Windowed aggregations
  • Stream-to-stream joins
  • Lookup enrichment
SELECT
  customer_id,
  COUNT(*) as order_count,
  SUM(total) as revenue
FROM orders
WHERE status = 'completed'
GROUP BY
  customer_id,
  TUMBLE(event_time, INTERVAL '1' HOUR)

Inline Inference

Run ML models and LLM calls directly on streaming events as a transform step. Every event can be enriched with predictions, classifications, or extracted entities before it reaches its destination.

  • Sentiment analysis on customer messages as they arrive
  • Fraud scoring on transactions in real time
  • Entity extraction from unstructured text fields
  • Classification and routing based on ML predictions
  • Embedding generation for vector search pipelines

Inference runs as part of your stream processing pipeline. No separate infrastructure, no batch jobs, no stale predictions.

Learn about AI Agents with Streamkap

Start processing streams today

Transform your data in real-time with SQL. No infrastructure to manage.

Start Free Trial