Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AniME: Adaptive Multi-Agent Planning for Long Animation Generation

Lisai Zhang, Baohan Xu, Siqian Yang, Mingyu Yin, Jing Liu, Chao Xu, Siqi Wang, Yidi Wu, Yuxin Hong, Zihao Zhang, Yanzhang Liang, Yudong Jiang

TL;DR

AniME addresses the challenge of automating long-form anime production with cross-stage consistency. It introduces a Director Agent orchestrating Specialized Agents via a Model Context Protocol to adaptively select control conditions and coordinate planning, memory, and quality control. Key contributions include a centralized Asset Memory Bank, explicit MCP-based tool selection for each sub-task, and structured inter-agent JSON communication with revision loops. The approach enables scalable, coherent generation from script to final video with character identity preservation and synchronized audio-visuals.

Abstract

We present AniME, a director-oriented multi-agent system for automated long-form anime production, covering the full workflow from a story to the final video. The director agent keeps a global memory for the whole workflow, and coordinates several downstream specialized agents. By integrating customized Model Context Protocol (MCP) with downstream model instruction, the specialized agent adaptively selects control conditions for diverse sub-tasks. AniME produces cinematic animation with consistent characters and synchronized audio visual elements, offering a scalable solution for AI-driven anime creation.

AniME: Adaptive Multi-Agent Planning for Long Animation Generation

TL;DR

AniME addresses the challenge of automating long-form anime production with cross-stage consistency. It introduces a Director Agent orchestrating Specialized Agents via a Model Context Protocol to adaptively select control conditions and coordinate planning, memory, and quality control. Key contributions include a centralized Asset Memory Bank, explicit MCP-based tool selection for each sub-task, and structured inter-agent JSON communication with revision loops. The approach enables scalable, coherent generation from script to final video with character identity preservation and synchronized audio-visuals.

Abstract

We present AniME, a director-oriented multi-agent system for automated long-form anime production, covering the full workflow from a story to the final video. The director agent keeps a global memory for the whole workflow, and coordinates several downstream specialized agents. By integrating customized Model Context Protocol (MCP) with downstream model instruction, the specialized agent adaptively selects control conditions for diverse sub-tasks. AniME produces cinematic animation with consistent characters and synchronized audio visual elements, offering a scalable solution for AI-driven anime creation.

Paper Structure

This paper contains 21 sections, 2 figures, 3 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Method
  3. Director Agent
  4. Workflow of the Director
  5. Asset Memory Management
  6. Specialized Agents and MCP Tools Selection
  7. Conclusion
  8. Method Details
  9. Director Agent
  10. Workflow of the Director
  11. Asset Memory Management
  12. Inter-Agent Communication Protocol
  13. Specialized Agents and MCP Toolsets
  14. Script and Storyboard Agent
  15. Character Designer
  16. ...and 6 more sections

Figures (2)

  • Figure 1: Brief Architecture of AniME
  • Figure 2: Storyboard keyframes for scene sceneYX01. Each shot shows the corresponding keyframe generated by the Storyboard Agent using selected tools (text-to-image, reference image, layout-guided generation).