Built to grow without breaking
Every piece of Transform Platform follows the Open/Closed principle — add capabilities without touching what already works.
Stream-First Engine
Records flow as a Kotlin coroutine Flow — no full load into memory. Process gigabyte files on a laptop without breaking a sweat.
Learn more →Plug-in Parsers
Add a new file format by implementing one interface and adding @Component. Spring auto-discovers it — zero changes to core.
Learn more →Errors Stay Local
Validation errors attach to individual records, not the pipeline. The stream never stops — bad records are quarantined, good ones ship.
Learn more →Dynamic Integrations
Add or rotate SFTP, Kafka, REST, and S3 connections at runtime via API. Hot-reload with AES-256 credential encryption — no restarts.
Learn more →Spec-Driven
FileSpec owns the schema. Parsers, correction rules, and validators all derive from it — a single source of truth per file format.
Learn more →Test-Friendly by Design
Open/Closed architecture: every parser, writer, and rule is independently testable. Kotest BDD specs ship with every module.
Learn more →Connect to anything, dynamically
Add or update client connections via API at runtime. Credentials are AES-256 encrypted, connectors hot-reload without service restarts.
Modular by design
Each module has a single responsibility. Deploy only what you need.
platform-commonShared models, interfaces & utilitiesplatform-corePipeline engine · parsers · writers · rulesplatform-apiREST API · FileSpec managementplatform-pipelineOrchestration · schedulingplatform-schedulerQuartz-backed job managementNew format in minutes
Implement one interface, drop one annotation. Spring discovers your parser or writer automatically — no changes to the registry, pipeline, or any existing code.
@Component
class NachaFileParser : FileParser {
override fun supports(format: FileFormat) =
format == FileFormat.NACHA
override fun parse(
stream: InputStream,
spec: FileSpec
): Flow<ParsedRecord> = flow {
// your parsing logic here
emit(ParsedRecord(...))
}
}Ready to transform your data pipeline?
Explore the docs, clone the repo, and have a working pipeline in under an hour.