A community-maintained Python framework for creating mathematical animations.
See the main README for some instructions on how to use this image.
Building the image
The docker image corresponding to the checked out version of the git repository can be built by running
docker build -t manimcommunity/manim:TAG -f docker/Dockerfile .
from the root directory of the repository.
Multi-platform builds are possible by running
docker buildx build --push --platform linux/arm64/v8,linux/amd64 --tag manimcommunity/manim:TAG -f docker/Dockerfile .
from the root directory of the repository.
To check a built image, render the verification scene (LaTeX, XeLaTeX, Pango, video encoding):
docker run --rm -v "$PWD/docker:/test:ro" manimcommunity/manim:TAG \
manim render -ql --media_dir /tmp/media /test/verify_image.py VerifyImage
Build notes
- The image is built via a multi-stage Dockerfile: TeX Live, the Python virtualenv and the runtime image are separate stages, and build dependencies are not carried into the runtime stage.
- TeX Live is pinned to the frozen final state of a past release (
TL_YEAR) from the TeX Live historic archive, with fallback mirrors (TL_REPOSITORIES). BumpTL_YEARonce the next release is frozen. TeX packages are listed intexlive-packages.txt. - Python dependencies are installed from
uv.lock. - CI builds each platform on a native runner and caches layers, so the TeX Live stage is usually not rebuilt.
Runtime notes
- The image does not include the
ffmpegCLI binary. - The default TeX installation is minimal and does not include
ctex.tlmgrinstalls further packages from the same frozen TeX Live release. - Headless OpenGL rendering relies on EGL/GL runtime libraries available in the image.