A universal iPad + Mac canvas app for visual data-science workflows, built on the SDSTK Swift data-science stack.
Drag-and-drop nodes — Data, Transform, Model, Visualize, Score, Export — connect into pipelines that run locally on-device. Workflows save as .sdstkflow documents; trained expert bundles export as .mbexpert packages.
- Xcode 26+ (iOS 26 / macOS 26 SDK)
- iOS 26.0+ or macOS 26.0+
- xcodegen (
brew install xcodegen) - SDSTK (SwiftSci) checked out as a sibling directory:
parent/
SDSTKStudio/ ← this repo
Downloads/
SwiftSci/ ← SDSTK source (required sibling path)
cd SDSTKStudio
xcodegen generate
open SDSTKStudio.xcodeprojGPU-accelerated nodes (Neural Network, Backend Benchmark) use mlx-swift. The Metal shader library is compiled natively by Xcode — no network access or manual steps required.
SDSTK Studio is dual-licensed:
Open-source: GNU Affero General Public License v3.0 or later (AGPL-3.0-or-later). Free to use, modify, and distribute under AGPL terms — including the requirement that any deployed service using this code make its source available.
Commercial: A commercial license is available for use cases where AGPL terms are not suitable (proprietary products, closed-source derivatives, SaaS without source disclosure). Contact jmelton@americancode.org for pricing and terms.
See NOTICE for the full copyright notice.