# Platform overview (/getting-started/overview) What Databaas is built from, and what each building block does for you. Databaas is a data platform assembled from established open-source projects rather than built as one closed product. Your data sits in open formats those projects define, not in anything proprietary to Databaas. This page explains what each part does for you; the rest of the docs explain how to use it. ## A platform of your own Your Databaas instance is dedicated to your organisation and runs where you choose: on-premise, or on your cloud of choice. You reach all of it from one portal with one login, and one set of permissions decides what you can open and what data you can read — in every tool. ## The building blocks | Layer | Built on | What it does for you | | ----- | -------- | -------------------- | | Tables | [Apache Iceberg](https://iceberg.apache.org/docs/latest/) | Your tables: typed schemas that can evolve, snapshots of every write, and time travel. An open standard governed by the Apache Software Foundation, readable by many engines. | | Arrays and files | [Zarr](https://zarr.readthedocs.io/en/stable/) and generic tables | Non-tabular data — forecast fields, rasters, sensor cubes, blobs — registered in the same catalog and governed by the same grants. | | Catalog | [Lakekeeper](https://docs.lakekeeper.io/) | Knows where every table lives and decides, per request, whether you may read or write it. No tool or person ever holds storage keys: access is granted to the data, not handed out as credentials. | | Notebooks | [JupyterHub](https://jupyterhub.readthedocs.io/en/stable/) and [marimo](https://docs.marimo.io/) | A private Python workspace per person, with DuckDB, Polars and pyiceberg ready to use. | | SQL | [Trino](https://trino.io/docs/current/) | Distributed SQL over the lakehouse. Your query runs next to the data and only the result comes back. | | In-process analytics | [DuckDB](https://duckdb.org/docs/stable/) and [Polars](https://docs.pola.rs/) | Fast SQL and dataframes inside your notebook or pipeline task, for data that fits in memory. | | Pipelines | [Apache Airflow](https://airflow.apache.org/docs/apache-airflow/stable/) | Scheduled and triggered pipelines, written in Python and kept in your own Git repositories. Works with [dlt](https://dlthub.com/docs/intro) for ingestion and [dbt](https://docs.getdbt.com/docs/introduction) for transformation. | | Dashboards | [Apache Superset](https://superset.apache.org/docs/intro/) | Charts, dashboards and the SQL Lab query editor. | | Identity | [Zitadel](https://zitadel.com/docs) | Single sign-on. You can sign in with your organisation's own account where your instance is connected to it. | ## How access works, in one paragraph Zitadel says who you are. Two separate kinds of permission then decide what you can do: **feature permissions** decide which tools you can open, and **data grants** in the catalog decide which warehouses, namespaces and tables you can read or change. Every query — from a notebook, SQL Lab, a dashboard or a pipeline — runs as the person or pipeline that started it, so a grant made once in the catalog applies in every tool. See [how access works](/access/how-access-works). ## Where to go next Sign in, find a table and read it. Warehouses, namespaces, engines and permissions. What the platform does not do, so you can plan around it.