#3636th of 102 in Image and videoFor image, video and design, image and design

heliosgen

Node-based desktop canvas for chaining image and video generation models

Releases and problems feed (RSS)

Stars
3.4k
License
no license
Last commit
Sep 2026
Last release
Sep 2026
Language
TypeScript
Stars, last 29 weeks+3,431
29 Apr 2026: 1 stars6 May 2026: 1 stars13 May 2026: 1 stars20 May 2026: 36 stars27 May 2026: 54 stars3 Jun 2026: 58 stars10 Jun 2026: 61 stars17 Jun 2026: 65 stars24 Jun 2026: 65 stars1 Jul 2026: 68 stars8 Jul 2026: 81 stars15 Jul 2026: 95 stars22 Jul 2026: 104 stars29 Jul 2026: 110 stars5 Aug 2026: 114 stars12 Aug 2026: 125 stars19 Aug 2026: 189 stars26 Aug 2026: 369 stars2 Sept 2026: 1,800 stars9 Sept 2026: 1,994 stars16 Sept 2026: 2,211 stars23 Sept 2026: 2,242 stars30 Sept 2026: 2,267 stars7 Oct 2026: 2,281 stars14 Oct 2026: 2,971 stars21 Oct 2026: 3,164 stars28 Oct 2026: 3,194 stars4 Nov 2026: 3,220 stars11 Nov 2026: 3,399 stars18 Nov 2026: 3,432 stars

1 on 29 Apr 20263,432 on 18 Nov 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

HeliosGen is a Tauri 2 desktop app with a Next.js UI that lets you wire prompts, reference images and models into pipelines on an infinite canvas. Generation is sent to the kie.ai API (GPT Image 2, Nano Banana, Seedream, Veo 3.1, Kling 3.0, Seedance 2.0 and others) using your own key. History, workflows and settings are stored in local SQLite, and media is saved to a local folder.

Who it is for: Creators building multi-step image and video generation workflows on macOS

Strengths

  • Parallel and sequential node pipelines with reference images on an infinite canvas
  • Covers many hosted image and video models through one kie.ai key
  • Workflows, history and settings kept in local SQLite, no account needed
  • Optional Codex CLI backend for GPT Image 2 using a ChatGPT subscription

Weaknesses

  • Generation runs on kie.ai, so it needs paid credits and network access
  • Prebuilt release is macOS Apple Silicon only; Windows and Linux need self-build
  • App is not code-signed; macOS Gatekeeper and Windows SmartScreen warnings appear
  • No support for local diffusion models or other API providers

What it needs

  • no GPU
  • Needs kie.ai API key, Codex CLI and uv (optional, for GPT Image 2 via Codex), Node 22+ and Rust (build from source only)
  • Models: GPT Image 2, Nano Banana, Seedream 5.0, Z-Image, Grok Imagine

What changed

From the release notes. No flag means the note does not say so, not that the upgrade is safe.

Also in Image and video

See all 102
Also in Image and video
RankProjectScore
1frigateLocal NVR with realtime object detection for IP cameras37.7k starsMIT94 out of 100
2toonflow-appInfinite-canvas tool for AI short drama and comic video production22.3k stars+3,812 (+21%) in 30 daysMIT89 out of 100
3MoneyPrinterTurboGenerates short videos from a topic, with script, footage, voiceover and subtitles144.7k stars+11,648 (+8.8%) in 30 daysMIT88 out of 100
#4InvokeAIWeb UI for diffusion image generation with inpainting canvas and node workflows29.1k stars+392 (+1.4%) in 30 daysApache-2.087 out of 100
#5ComfyUINode-graph editor for running diffusion image, video, audio and 3D workflows147.7k stars+7,765 (+5.6%) in 30 daysGPL-3.086 out of 100