SegFault42/HeliosGen

What it solves

HeliosGen addresses the complexity of creating multi-step AI image and video generation processes by providing a visual, node-based interface. It eliminates the need for manual prompt chaining and vendor lock-in by allowing users to build reusable pipelines and self-host their infrastructure.

How it works

Users utilize an infinite canvas to drag and connect nodes that represent prompts, different AI models, reference images, and automation steps. These nodes can be executed in parallel or sequentially to create complex generation workflows. The system integrates with Kie.ai for model access and can optionally use the Codex CLI for specific OpenAI image generation.

Who it’s for

It is designed for AI creators and developers who want to build sophisticated image and video generation pipelines without writing code for every step, as well as those who prefer self-hosted tools over subscription-based platforms.

Highlights

• Infinite Node Canvas: A visual drag-and-connect system for building AI workflows.
• Multi-Model Support: Integrates various image and video models from providers like Google, OpenAI, Kling, and X.
• Flexible Execution: Supports both parallel and sequential pipeline execution.
• Self-Hostable: Can be deployed on platforms like Vercel, Railway, or Render using Supabase and Cloudflare R2.
• Reference Image Support: Allows the use of images to guide the generation process.