Developers
A direct path to clone, inspect, build, and benchmark.
The public repository provides architecture, SDK contracts, reproducible evidence, evaluation paths, and enough surface area to understand the plugin model.
Quick start
Build the app and run a benchmark profile.
The first developer experience should be concrete: build the shadow jar, then run a known benchmark row from the published matrix.
PowerShell
.\gradlew.bat --no-daemon :streamkernel-app:shadowJar
.\test-java-runner.ps1 -MatrixFile .\benchmark-runs\tests.csv -SingleTest streamkernel_kafka_at_least_once_baseline_10m Linux / macOS
./gradlew --no-daemon :streamkernel-app:shadowJar
./test-java-runner.sh -MatrixFile ./benchmark-runs/tests.csv -SingleTest streamkernel_kafka_at_least_once_baseline_10mEvaluation map
Repo areas the website should send engineers toward.
Plugin surfaces
- Source plugins for Kafka, Pulsar, REST, and custom event ingress.
- Policy plugins for governance before delivery, including OPA-backed deny paths.
- Transformer chains for in-process enrichment.
- Sink plugins for Kafka, MongoDB Vector, Postgres, Postgres Vector, Delta Lake, Snowflake, and custom targets.
- Metrics providers for Prometheus and OpenTelemetry runtime evidence.
Plugin-owned extensibility
Bring plugins without losing the auditable kernel boundary.
The public SPI supports evaluation and extension. Commercial engagement covers protected AI implementation details, production packaging, support, redistribution, and OEM embedding.
Commercial path
Need a guided technical evaluation?
Use the commercial path for benchmark review, protected AI builds, plugin guidance, and production readiness planning.