About Dataglot

What Dataglot is

Dataglot is an open-source, Rust-native federated SQL query engine with governance enforced in the query plan. It speaks the PostgreSQL wire protocol, so every client, driver, and BI tool that talks to Postgres talks to Dataglot — and it queries the databases you already run, in place: PostgreSQL, MySQL, Oracle, Snowflake, object storage, and Apache Iceberg.

The differentiator is where governance happens. Column masks and row filters are compiled into the physical query plan as typed expressions, pushed into the scan itself. A masked column is never fetched from storage. There is no post-processing step to bypass, no SQL-rewrite proxy to route around — the policy is structural.

Why it exists

Regulated enterprises — banks, insurers, healthcare providers — sit on data they are not allowed to move and are audited on how it is accessed. The tools they use to define policy (data catalogs and governance platforms) are mature. The layer that enforces those definitions at query time, across every source and every client, is what's been missing. Dataglot is that layer: your governance platform defines what PII means in your organization; Dataglot is where that definition becomes a query-time guarantee.

What Dataglot is not

Dataglot is not a SaaS data-integration tool. Products like Airbyte, Fivetran, and Hightouch exist and are excellent at moving data. Dataglot is the governance-enforcing query layer above the data you keep in place, and it partners with — never competes with — the governance platforms where policy is authored.

Who builds it

Dataglot is built by Peaka and developed in the open at github.com/dataglotai/dataglot under the Apache-2.0 license. It stands on the shoulders of Apache DataFusion, Apache Arrow, Apache Iceberg, and Apache Ballista.

We didn't invent a paradigm. We built cleanly on the one that won.

Get in touch

• Issues and discussions: GitHub
• Security reports: see SECURITY.md