Command-line tool and Python library that turns a whiteboard-style image into a CPU-rendered hand-drawn reveal MP4, optionally paced to narration.
3b1b/manim
ManimGL is the original 3Blue1Brown engine for writing precise, code-driven animations used to make explanatory math videos.
- Stars
- 95,326
- Forks
- 7,813
- Open issues
- 505
- Last commit
- last month
What is it?
ManimGL is a Python engine for programmatic animation, built by the author of 3Blue1Brown to produce the math explainer videos on that channel. Scenes are written as Python classes, and objects such as shapes, equations, graphs and 3D elements are animated through code, which makes timing and geometry exact and repeatable.
The project is the original 3b1b version of Manim, distinct from the community fork maintained as ManimCommunity. It is installed from PyPI as the package manimgl and run with the manimgl command, which can open an interactive preview window or render video files. Rendering relies on FFmpeg and OpenGL, and LaTeX is optional for typeset math.
The repository also includes a CLI with flags for writing output files, skipping to a final frame and jumping to a specific animation, along with a custom_config.yml for output paths and render defaults.
Key features
- 01Defines animations and scenes in Python code for exact, repeatable timing and geometry.
- 02Previews scenes in an interactive window, or renders them to video files.
- 03Supports LaTeX for typeset mathematical formulas.
- 04Offers CLI flags to save output, jump to a final frame or a chosen animation.
- 05Configurable output paths and render defaults via custom_config.yml.
Use cases
- Producing explanatory math or science videos with code-controlled animation.
- Visualizing equations, graphs and geometric transformations step by step.
- Prototyping animated diagrams that need precise, reproducible timing.
Quick start
- 1
Install ManimGL from PyPI
pip install manimgl - 2
Run the example scene to check it works
manimgl
Things to know
- Requires FFmpeg, OpenGL and optionally LaTeX; Linux also needs Pango headers.
- The package name manimgl differs from the community edition, so installing the wrong version can cause conflicts.
- Needs Python 3.10 or higher.
Topics
FAQ
- What is manim?
- ManimGL is the original 3Blue1Brown engine for writing precise, code-driven animations used to make explanatory math videos. ManimGL is a Python engine for programmatic animation, built by the author of 3Blue1Brown to produce the math explainer videos on that channel. Scenes are written as Python classes, and objects such as shapes, equations, graphs and 3D elements are animated through code, which makes timing and geometry exact and repeatable.
- Is manim free to use? What license does it have?
- manim is open source under the MIT license. Check the license terms before using it in commercial products.
- Is manim still maintained?
- Updated within 6 months. The last commit to manim was on Sep 9, 2026.
- What can I use manim for?
- Producing explanatory math or science videos with code-controlled animation. Visualizing equations, graphs and geometric transformations step by step. Prototyping animated diagrams that need precise, reproducible timing.
- How do I install manim?
- Install ManimGL from PyPI: pip install manimgl Run the example scene to check it works: manimgl
- Where can I get manim?
- The source code and installation instructions are on GitHub: https://github.com/3b1b/manim
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A Python and JavaScript engine that converts video into ASCII or pixel-block frames and streams them to an HTML5 Canvas player in real time.
An open-source framework that turns HTML, CSS, media and seekable animations into deterministic MP4 videos, built for AI coding agents.
Open-source Rust toolkit for logging, visualizing, querying and streaming multimodal robotics and sensor data over time.
Converts static diagram images into editable DrawIO (XML) files using SAM 3 segmentation, OCR, and LaTeX formula recognition.
A set of Agent Skills that teach AI coding agents such as Claude Code, Codex and Cursor best practices for building Remotion video projects.