Overview
Build DAGs of source, transform, and sink nodes on a canvas; write SQL (DataFusion) or Python (Polars) transforms; and see live data previews as you go. Armillary ships as a single Rust binary with an embedded browser UI, with prebuilt Docker images on every release.
Highlights
- Visual DAG canvas — wire up sources, transforms, and sinks by hand.
- SQL or Python — DataFusion SQL or Polars Python per transform node.
- Live previews — see the data flowing through each node as you build.
- Single binary — one Rust executable with the UI embedded; no extra services.
Install
# Pull and run (also on GHCR: ghcr.io/cache-has/armillary)
docker pull cachehorizon/armillary:latest
docker run -p 8080:8080 -v armillary-data:/data cachehorizon/armillary:latest Then open http://localhost:8080.