HeliosGen

Build AI image & video pipelines visually.
Chain prompts, models, reference images, and automations on an infinite canvas.

HeliosGen is a desktop app. Grab the latest build for your OS from the releases page — no account, no server, no cloud setup:

Only the builds actually attached to the latest release are available. macOS is published today. Tauri can't cross-compile, so Windows and Linux builds need someone on those platforms — if you can run npm run desktop:build on Windows or Linux (see the Build from source section below), please open a PR or attach the artifacts to an issue and we'll add them to the release.

The app is not code-signed yet:

• macOS — right-click the app → Open (once), or run xattr -cr /Applications/HeliosGen.app.
• Windows — SmartScreen: More info → Run anyway.

1. Launch HeliosGen.
2. Open Settings → API Keys and paste your kie.ai API key.
3. Start generating.

Everything stays on your machine. Generations, uploads, folders, workflows and settings live in a local database; media is saved to a local folder:

Delete that folder to reset the app.

Image generation example

Workflow generation example

JSON prompt preview

Prompt assistant interface

HeliosGen is a free & open source visual AI workflow builder for image and video generation.

Build reusable AI pipelines with:

• infinite node-based workflows,
• multi-model generation,
• reference images,
• automation chains,
• all running 100% locally on your machine.

No subscriptions.
No disappearing credits.
No vendor lock-in.
No cloud, no accounts — just a local app and your own kie.ai key.

HeliosGen now works with kie.ai.

All credits are purchased directly on your own account and never expire.

That means:

• no monthly reset,
• no lost credits,
• no subscription lock-in,
• and full ownership of your usage.

You only pay for what you generate.

• Infinite node-based canvas
• AI image & video generation
• Drag-and-connect workflow system
• Multi-model pipelines
• Reference image support
• Parallel & sequential pipeline execution
• Real-time generation history
• 100% local — your data never leaves your machine
• Bring your own kie.ai key
• Modern responsive UI

• GPT Image 2 (OpenAI)
• Nano Banana / Nano Banana 2 / Nano Banana 2 Lite / Nano Banana Pro (Google)
• Seedream 5.0 Lite / Pro (Seedream)
• Z-Image (Z-AI)
• Grok Imagine (X)

• Veo 3.1 Lite / Fast / Quality, Gemini Omni Video (Google)
• Kling 3.0, Kling 3.0 Turbo, Motion Control 2.6 / 3.0 (Kling)
• Seedance 2.0 / Fast / Mini (Bytedance)
• Grok Imagine, Grok Imagine 1.5 preview (X)
• HappyHorse (Alibaba)

More models are coming.

Instead of routing GPT Image 2 through kie.ai credits, HeliosGen can generate through your own ChatGPT Codex subscription via codex-imagegen-cli. The desktop app picks up codex from your PATH automatically; if it's missing, the feature just shows NOT CONFIGURED and everything else keeps working.

Requirements:

• A ChatGPT Plus/Pro/Team/Enterprise account with Codex access
• codex CLI installed on your machine
• uv (Python package manager)

# macOS
brew install codex
# or, cross-platform
npm install -g @openai/codex

git clone https://github.com/jdmnk/codex-imagegen-cli.git
cd codex-imagegen-cli
uv sync --dev
uv tool install -e .

This installs the codex-imagegen binary — make sure it's on your PATH.

Either:

• run codex login in a terminal (opens a browser to sign in), or
• open the app → Settings → API Keys → Codex CLI → Connect Codex, which walks you through a device-code login — visit the printed URL and enter the code, no terminal needed.

⚠️ Starting a new login (either way) immediately invalidates any existing session on that machine — the CLI clears old credentials the moment a login attempt begins, whether or not it's ever completed. Only start one when the status badge below shows NOT CONFIGURED.

In Settings → Image Models, set GPT Image 2's provider toggle to Codex CLI. The status badge in Settings → API Keys shows READY once both the CLI and login are in place.

Prefer to build it yourself, or need a platform that isn't on the releases page yet? The whole app builds from this repo.

Tauri does not cross-compile — build on the OS you want to target. Run npm run desktop:build on a Mac for the macOS build, on Windows for Windows, on Linux for Linux.

Want to help ship Windows / Linux builds? Build on that OS and send the artifacts (PR or issue attachment) — they'll be added to the next release, with credit.

Platform-specific system deps:

• macOS — Xcode Command Line Tools: xcode-select --install
• Windows — Microsoft C++ Build Tools (Desktop development with C++) and WebView2 (preinstalled on Windows 11)
• Linux — webkit2gtk-4.1, librsvg2, build-essential, curl, wget, file, libssl-dev, libayatana-appindicator3-dev (Debian/Ubuntu package names; see the Tauri prerequisites page for other distros)

git clone https://github.com/SegFault42/HeliosGen
cd HeliosGen
npm install
npm run desktop:build

Artifacts land in src-tauri/target/release/bundle/:

The macOS build is unsigned — on first launch Gatekeeper blocks it. Right-click → Open, or xattr -cr "src-tauri/target/release/bundle/macos/HeliosGen.app".

npm run desktop:dev

Runs next dev and tauri dev together. The first run compiles the Rust shell (~1–2 min).

See DESKTOP.md for architecture, data locations, and signing & notarization.

Contributions are welcome.

If you find a bug, have an idea, or want to improve HeliosGen:

• Open an issue
• Submit a pull request
• Share feedback or feature requests

All contributions are appreciated.

MIT License

Built for creators building the future of AI workflows.