AI Video Generator for Game Devs & VTubers

Game developers, comic creators, and VTubers have the worlds, characters, and stories. But turning those into video has always meant hiring a film crew, spending weeks in editing software, or learning complex motion graphics tools. It's slow and expensive, especially for indie creators working on tight budgets.

I built an AI-powered video generation pipeline on Melius that takes a creator's vision and turns it into finished video, node by node, using AI agents that handle each step of the production process.

Sign up for a Melius account and create a new project

Brief the Melius AI agent about your video concept: characters, scenes, tone, style

The AI agent builds the full production workflow automatically on the canvas

Each node in the workflow handles a specific task: image generation, scene composition, video rendering, and stitching

Review the output at each stage, tweak anything that doesn't match your vision

The stitch utility node combines all video segments into the final long-form video

The entire pipeline runs on the Melius canvas with no manual editing software required.

Seedance 2.0 for video generation

Claude Sonnet 4.5 for script and brief processing

GPT Image 2 for image generation and scene creation

Saves time: What used to take weeks of production and editing now happens in a fraction of the time

Cuts costs: No film crew, no expensive editing software, no motion graphics freelancers

Creator-friendly: The canvas-based workflow is lightweight and intuitive, even for non-technical users

Full creative control: Review and modify output at every node before moving to the next stage

Built for the Contra x Melius Challenge.