AI lip sync for Blender, from audio to animated mouth in minutes.
The biggest update yet. Tap any tab to explore the new features.
Whisper and Allosaurus now side by side. Whisper estimates phonemes from speech-to-text. Allosaurus detects them directly from audio acoustics, more precise, and supports 2,000+ languages.
Only the dominant viseme gets set to 1. Everything else drops to 0. A hard switch between mouth shapes that gives a clean, snappy 2D-style animation feel, great for stylized or cartoon characters.
Available for Shape Keys and Bone modes. Can be toggled per-project in Processing Options.
A dedicated panel handles Butterworth filter smoothing for softer keyframe curves, and animation transfer (copy, paste) between compatible characters.
Zero extra setup needed. Apply smoothing and transfer directly after generating your animation.
Randomized blink animations generated automatically across all animation modes. No manual keyframing, no extra setup.
The Eye Blinks panel lets you control blink frequency and intensity to match your character's personality.
After generating, navigate and inspect each phoneme keyframe individually. Replace, reassign, or remove specific mouth shapes without re-processing the full audio.
Available for Shape Keys mode only.
Shape keys, bone actions, or ARKit blendshapes, Panim adapts to whatever your character uses.
Panim is made for Blender artists who need fast lip sync without rebuilding their rig workflow.
Generate mouth animation from audio for shape keys, bone actions, or ARKit rigs without changing how your character is built.
Batch process multiple characters and audio clips with frame markers, built for scenes with more than one speaker.
Use smooth blending for natural motion or Binary Viseme mode for clean, readable cartoon-style mouth shapes.
From multilingual audio to production-ready bone rigs, Panim gives you the tools to do it right.
Allosaurus uses universal IPA phoneme recognition across 2,000+ languages, no language configuration needed. Whisper supports dozens of the most common languages with a configurable language setting.
Switch engines freely depending on the audio you're working with.
Each phoneme group points to the specific blend shapes or bone actions in your rig. Set it up once and reuse on every audio file.
Supports Shape Key, Bone, and ARKit modes with auto-detection for standard naming or manual override for any custom rig.
Add multiple characters, assign individual audio clips with frame markers, and process them all in a single click. Built for scenes with more than one speaking character.
Built for Blender character rigs: local AI phoneme detection, Shape Keys, Bones, ARKit, batch dialogue, and post-processing tools in one workflow.
Direct acoustic phoneme detection from audio. Supports 2000+ languages out of the box.
Hard switch between mouth shapes. Only the dominant viseme activates at full value.
Butterworth smoothing, animation transfer, and delete controls in one panel.
Randomized blink generation with auto-detect across all animation modes.
Navigate and replace any phoneme keyframe after baking the animation.
Visible bones only by default. Much faster performance on complex rigs.
Icon-only toolbar with dedicated panels for a cleaner workspace.
Dependencies survive Blender restart. scipy installs safely with --no-deps.
From raw audio to fully animated mouth, done in minutes, or even seconds.
Import MP3, WAV, OGG, FLAC, M4A, or AAC. Works in any language with no extra configuration.
Panim detects phonemes locally using your chosen engine and maps them to your rig's visemes.
Shape keys or bone actions get keyframed automatically. Refine individual frames as needed.
Panim runs entirely on your machine. No cloud service, no internet required after the initial model download.
Mac support is limited and not actively tested. Windows is the primary supported platform.
Windows requires a one-time Blender restart after installing dependencies.
Internet is only required once per engine to download the AI model. After that, everything runs locally.
Refinement Panel is available for Shape Keys mode only, not for Bone Actions or ARKit.
"character_02" by BuuMya is licensed under Creative Commons Attribution (Sketchfab).
"2D mouth with shape keys phonemes" by myname, from BlendSwap.com (BlendSwap).
"Ellie - Character Rig" from Blender Studio (Character Link).
"Victoria - Character Rig" from Blender Studio (Character Link) (Temporarily displayed in the video).
To allow characters to be used for demonstration of the tool.
The characters, music, and other files used in demonstrations and examples are not included with this addon. If you want to test with similar assets, you can download them from the sources listed in the "Thanks to" section above.
Thank you!