Open-source SQL ELT platform

Build and control data pipelines without losing visibility

Belt helps data teams and business users design ELT flows, run process chains, monitor every request and batch, and safely recover data when something goes wrong

Apply for free pilot

MVP available for pilot deployments. PostgreSQL, ClickHouse and Python supported. Addition connectors at the future.

Source PostgreSQL, ClickHouse, Python and so on

SQL Transform Analytics can write SQL transformations

Target Request + batch tracking, data management and lineage

Flow run

Batch error

Data lineage

Reload

The gap

Data loading is easy
Data recovery is not

Many teams can create pipelines, but later struggle to answer simple operational questions: which request loaded these rows, which batch failed, what can be safely deleted, and how to reload only the broken part. Belt is focused on this operational layer.

Visual ELT flows

Create integration and transformation flows visually, while keeping SQL clear and readable.

Request and batch control

Track each run by technical request and batch identifiers, down to the loaded rows.

Monitoring and recovery

Inspect chain runs, flow runs and batch status. Delete or reload selected requests when needed.

Lineage view

See how sources, transformations and targets are connected across flows.

SQL-first transformations

Write business SQL. Belt adds technical metadata automatically during loading.

Simple deployment

Run Belt with Docker on your own server, close to your databases and internal systems.

Who needs it

For small and medium data teams that need control, not platform bureaucracy

Teams that need visible, recoverable, request-based ELT
BI teams building internal data marts
Business users and analysts who know SQL
Companies moving data into DWH
Teams that need ELT, scheduling, monitoring and lineage in one tool

Pilot program

Try Belt on real pipelines

We are looking for 2–3 companies to run Belt on real ELT processes. Pilot teams get free deployment, setup assistance and direct support in exchange for feedback and a short case study if the pilot is successful.