SDCBench is a desktop tool for diagramming a data model and handing it to a data modeler to finish in SDCStudio. A domain expert reuses published components that already mean what their data means, sketches a new one only when needed, and says in plain language what each piece is, without ever learning the underlying Semantic Data Charter reference model.
You sign in with your SDCStudio API key, then:
- Search the published library and reuse components by dragging them onto a canvas (reuse is the default, and the cheap path).
- Assemble them into a document tree (Model → Group → Fields) on a Google Blockly canvas, where blocks only snap where the reference model allows.
- Sketch a new component only when nothing fits: pick a plain data type (Text, Integer, Decimal, Date, Code, …) and write a plain-language requirement.
- Send to SDCStudio: it creates a draft data model in your project, referencing reused components and creating new ones as drafts.
A data modeler then finalizes the constraints, units, reference ranges, and semantic bindings, and publishes, in SDCStudio.
| SDCBench (domain expert) | SDCStudio (data modeler) | |
|---|---|---|
| Does | reuse published components, sketch new ones, structure the tree, state requirements in plain language | constraints, units, reference ranges, semantic binding, publish |
| Produces | a draft model (reuse by ct_id + new drafts) with per-component requirements |
the finished, published model |
- An SDCStudio account and API key (create one from the sign-in screen).
- A running SDCStudio instance (the public one by
default; configurable via
SDCSTUDIO_BASE_URL).
SDCBench is a Tauri app (Rust core + a Vite/vanilla-JS frontend). You need Rust and Node.js.
cd app
npm install
npx tauri dev # run in development
npx tauri build # build a release bundle (Linux AppImage)- Windows: see
docs/BUILDING-WINDOWS.md. - macOS is not yet supported.
CI (GitHub Actions) builds the Linux AppImage and a Windows installer on tags.
Beta installers are not yet code-signed; Windows code signing is wired into CI and
turns on once configured. See docs/CI-SIGNING.md.
docs/USER-GUIDE.md— the end-to-end walkthrough (also in-app via the Help button).docs/COMPOSITION-MODEL.md— how the canvas's drop-validity is derived from the reference model.docs/PHASE-2-PRD.md— the product design.
Apache License 2.0. Copyright 2026 Axius SDC, Inc. See LICENSE.