HeliosGen is a desktop application for constructing AI‑driven image and video generation workflows on an infinite node‑based canvas. Built with a Tauri 2 shell (Rust) and a bundled Next.js + React + TypeScript sidecar, it stores workflows, settings and generation history in a local SQLite database while saving media to a user‑specific folder (~/Library/Application Support/cash.sdd.helios.desktop on macOS, %APPDATA%\cash.sdd.helios.desktop on Windows, and ~/.local/share/cash.sdd.helios.desktop on Linux). The app connects to the kie.ai service for models such as GPT Image 2, Nano Banana families, Seedream 5.0 Lite/Pro, Z‑Image, Grok Imagine, Veo 3.1 variants, Kling 3.0/Turbo, Seedance 2.0, and HappyHorse, and can also route GPT Image 2 through a locally installed OpenAI Codex CLI when a ChatGPT subscription with Codex access is present. Workflows support drag‑and‑connect nodes, multi‑model pipelines, reference images, and parallel or sequential execution, all running entirely offline with no accounts, subscriptions, or cloud dependencies. Building the app requires Node 22+, Rust, and Tauri system dependencies; the command npm run desktop:build produces platform‑specific bundles (DMG for macOS, MSI/NSIS for Windows, DEB/RPM/AppImage for Linux), though only the macOS build is currently published in the releases, with Windows and Linux builds awaiting community contributions. The software is released under the MIT license.

Read original