Lipkit is a Blender add-on from codeideal that generates lip sync directly from audio. It draws on the Rhubarb phoneme engine and adds keyframes to a controller, cutting out much of the manual keyframing that character dialogue usually requires. Audio can come from the Blender Video Editor or from external sound files.
The add-on works for both 2D and 3D. It drives Grease Pencil layer opacity for drawn mouth shapes and shape keys for 3D models, and its built-in controller objects make individual mouth poses easy to tweak. Phoneme presets are fully mappable, so any shape key or Grease Pencil layer can be linked to any mouth pose. A single project can hold multiple characters and multiple voice lines at once.
Built with extendability in mind, Lipkit is set up to support additional phoneme engines in future releases. It is compatible with Blender 4.4 through 5.1 and is distributed under a GPL license.