Duckle is a free, open-source, local-first Data Studio that runs on your laptop!

Duckle is a free, open-source, local-first Data Studio that runs on your laptop: build pipelines on a visual canvas, run them on DuckDB, ship them as a
single binary. No cloud, no account, no telemetry. Your data never leaves your machine.

The latest build (v0.3.0) makes dbt a near-instant, cross-system part of the Duckle Canvas:

- dbt is now supported and dbt Fusion is now the default. A Rust dbt engine: warm project parse/build is ~45 ms, versus the multi-second Python import floor of dbt Core (which is on as an automatic fallback).
- Multi-source dbt. One dbt build reads several wired sources at once (Postgres + MySQL + CSV + Parquet), each materialized as a real table and modeled
through dbt sources. A Customer 360 demo runs 6 sources across 4 system types into 1 dbt build and out to 4 sinks in 4,382 ms.
- Free, self-provisioning. The dbt engine downloads and sets itself up on first launch. No Python setup, no separate install, $0.
- JSON Records-path. Unnest nested REST envelopes (like data or response.records) into real columns.
- Native brand icons + type-to-add. Every source, sink and SaaS connector wears its real logo on the canvas; start typing to fuzzy-search and drop any
connector.
- Production ops. Structured error taxonomy, OpenMetrics export(<workspace>/runs/*.json), backfill and watermark controls, and a Runs history tab.
- Right-click the pipeline, choose Build, and it compiles into a self-contained executable, including DuckDB and it's necessary extensions.
Just copy that file to a server.

Single binary. Engines download on first launch. No installer, no JVM, no control plane. Swap the binary in place and your workspace + engine cache are
untouched.

Repository: https://github.com/SouravRoy-ETL/duckle
Download + full changelog: https://github.com/SouravRoy-ETL/duckle/releases/tag/v0.4.0