Orelake

In development · October 2026

Send raw data.
Ask anything.

Orelake is a data platform for teams that have data but no data team. You send records as they are. Orelake works out the schema, builds and maintains the tables and rollups from how the data is actually used, and answers questions in one place.

How it is meant to work

1. 1

   Send

   Push raw events and records over a Kafka-compatible endpoint or in files. No schema design up front.

2. 2

   Shape

   The system infers schemas, writes open Apache Iceberg tables, and builds rollups from real query patterns, with guardrails on every change.

3. 3

   Ask

   Query in SQL or plain language and get tables, charts and written answers. Reports and sharing come from the same place.

What we are building on

Rust · Apache Arrow · Apache DataFusion · Apache Iceberg · Parquet · S3-compatible storage · Kafka protocol

Open formats throughout: your data stays in Iceberg tables you can read with other engines.

Where we are

• Now: foundation work. Build and test tooling, a local stack, a first write-and-query path through Iceberg, and benchmark and correctness harnesses.
• Next: ingest, ask and guardrails: the first version someone outside the team can try.
• Not yet: there is no product to sign up for today. This page will say so when there is.

Contact

Questions or interest in early access: hello@orelake.com