Skip to main content
DAESAIX
In development · Early access

DAESAIX DataStack

Your data platform, in your infrastructure.

DataStack connects your databases, object stores, files and APIs, runs pipelines on DuckDB or Spark, and lands data as Parquet or Apache Iceberg tables in storage you control. Every run records lineage automatically, so you always know which source produced which table, through which pipeline.

How data moves

  1. Sources
  2. Connectors
  3. DuckDB or Spark
  4. Parquet / Iceberg
  5. MinIO / S3
  6. Assets + lineage

The Problem

Getting data ready shouldn't mean giving it away.

Most teams choose between months of custom pipeline code or a SaaS tool that sends their data through someone else's cloud and bills by volume. Either way, lineage and governance get bolted on later—if at all.

Built for

  • Data teams tired of brittle hand-coded pipelines
  • Organizations that need data to stay inside their own perimeter
  • Companies that want to avoid per-connector and per-gigabyte pricing

What Works Today

Capabilities

Connector framework

Relational databases, object stores, files and APIs through one extensible connector registry.

Pluggable compute

DuckDB for lightweight workloads and Spark (local or standalone) for heavier batch jobs—behind one engine interface.

Lakehouse-native output

Parquet files or managed Apache Iceberg tables with partitioning, safe schema evolution, snapshots and time travel.

Automatic lineage

Source → pipeline → engine → format → storage is recorded on every run, including Iceberg snapshot and schema IDs.

Storage you control

Local disk, MinIO or S3. Metadata and data stay in your infrastructure—data residency by default.

UI and API

A web UI with a data explorer, plus a full API, so pipelines can be built by hand or driven from your own tooling.

Deployment

  • Self-hosted with Docker Compose
  • PostgreSQL metadata store, Redis, MinIO
  • Starts on a laptop, grows toward VMs and clusters
PythonFastAPINext.jsDuckDBApache SparkApache IcebergPostgreSQLMinIO

On the roadmap

  • Automatic DuckDB / Spark selection by workload size
  • Delta Lake table format and REST / Glue catalogs
  • Data masking and privacy policies for sensitive fields
  • Kubernetes deployment

See DAESAIX DataStack on your own landscape.

Early-access teams get a guided setup and a direct line to the people building it. Prefer us to run it for you? It's part of our data engineering engagements.