The process of creating animations with Manim is powerful but time-consuming, requiring precise scripting and mathematical transformations. While tools like Adobe After Effects offer intuitive interfaces, they lack Manim’s mathematically precise control.
This project aims to bridge the gap between scripting-based animation and creative sketching by introducing an AI-assisted workflow. The software will allow users to sketch vector-based shapes and provide text-based prompts, which an AI model will convert into Manim animation scripts.
Initially, the proof-of-concept (PoC) version will focus on: - Targeting experienced Manim programmers, ensuring they can refine AI-generated scripts. - Generating raw Manim code only, without real-time preview adjustments. - Providing a simple sketchpad interface, allowing users to draw vector shapes that serve as a basis for AI-generated animations.
The proof-of-concept version of the AI-supported animation software will focus on core functionality and essential usability. Features are prioritized using MoSCoW categorization.
Must-Have (M): - A vector-based sketchpad allowing users to draw basic shapes (lines, circles, polygons). - A text prompt input for users to describe animation behavior. - AI-powered Manim code generation based on sketches and prompts. - A code editor where users can refine the generated Manim script. - Export functionality to save the generated Manim script for execution.
Should-Have (S): - A basic UI similar to Adobe Illustrator/Inkscape, focused on simplicity. - Support for layers or groups to organize vector elements. - Undo/Redo functionality in the sketchpad.
Could-Have ©: - SVG import support for predefined vector shapes. - Manim preview integration (but not real-time rendering). - AI-based refinements, such as auto-smoothing of paths.
Won’t-Have (W) (for PoC phase): - Real-time animation previews. - Advanced animation timelines or keyframing. - Deep learning-based motion inference.