HKUDS/ViMax

"ViMax: Agentic Video Generation (Director, Screenwriter, Producer, and Video Generator All-in-One)"

What it solves

ViMax addresses the limitations of current AI video tools, which are often restricted to short clips, suffer from inconsistent characters and scenes, and lack narrative structure. It automates the complex production pipeline—including scriptwriting, storyboarding, and shot design—to enable the creation of high-quality, long-form videos from simple concepts or existing scripts.

How it works

ViMax acts as an all-in-one agentic framework that orchestrates several creative roles (Director, Screenwriter, Producer, and Video Generator). It connects narrative planning with visual generation through several specialized workflows:

  • Idea2Video: Converts a short concept into a full production (stories, characters, scripts, and storyboards).
  • Script2Video: Transforms a detailed screenplay into a multi-scene video while preserving creative intent.
  • Novel2Video: Adapts long-form fiction into episodic visual narratives using narrative compression and character tracking.
  • AutoCameo: Integrates a reference photo of a person or pet to maintain a consistent appearance throughout the story.

The system uses a combination of LLMs for planning and specialized image/video generators for rendering. It includes an "Agent Loop" and a Web UI for interactive planning, revision, and rendering control.

Who it’s for

It is designed for independent creators and rapid prototypers who want to produce professional-grade video content without needing a full team of specialists in cinematography and visual storytelling.

Highlights

  • End-to-End Automation: Handles everything from the initial prompt to the final assembled video.
  • Visual Consistency: Coordinates references and first frames to ensure characters and environments remain stable across shots.
  • Interactive Control: Features a Web UI and TUI for discussing ideas with the agent, reviewing artifacts, and managing projects.
  • Parallelized Generation: Generates compatible shots and media assets concurrently to speed up production.
  • Flexible Input: Supports concepts, explicit scripts, or long-form novels as starting points.

Related

  • Project
    Anil-matcha/Open-AI-Micro-Drama-Generator

    An agentic AI pipeline that transforms a simple idea or script into a complete cinematic micro-drama video by automating story development, character design, and video production.

  • Project
    HKUDS/VideoAgent

    An all-in-one agentic framework for understanding, editing, and generating creative videos through natural language prompts.

  • Project
    ChrisChen667788/wind-comic

    Wind Comic is a web‑based, multi‑agent AI studio that converts a one‑sentence prompt into a finished short‑form drama (script, storyboard, voice‑over, lip‑sync, music, MP4). It chains specialised agents (writer, director, character designer, video engine, TTS, lip‑sync, editor) under a real‑time collaborative timeline, supports bring‑your‑own LLMs, plugs in 12+ image/video generators, provides cost‑guarding and a template market, and runs on Next.js 16 + Node 20 with SQLite/PostgreSQL storage.

  • Project
    0xsline/StoryGen-Atelier

    An AI-assisted tool that generates storyboards and stitches them into coherent videos using Gemini and Vertex AI Veo.

  • Project
    ItusiAI/Open-Magiviz

    An AI-powered video creation platform that automates the workflow from script generation and character design to storyboard creation and final video rendering.