Codex and Claude Code plugin for adding MotionPNGTuber / MotionPNGTuber_UI style talking characters and Japanese TTS narration to Remotion or HyperFrames videos.

The bundled skill covers:

• Rendering MotionPNGTuber mouth sprites as a frame-driven canvas overlay from mouth_track.json in either Remotion or HyperFrames.
• Preserving animated mouthless body video or synchronized extracted frames.
• Falling back to the bundled default PNGTuber model in assets/default-pngtuber/nike_loop_fix when no model is specified.
• Generating VOICEVOX or AivisSpeech narration through a local VOICEVOX-compatible HTTP API.
• Validating visual sync, transparent compositing, and final MP4 output.

Add this repository as a Codex plugin marketplace:

codex plugin marketplace add https://github.com/tegnike/remotion-motionpngtuber.git

Then restart Codex, open the plugin directory, select the MotionPNGTuber for Remotion and HyperFrames marketplace, and install the plugin.

For local development from this folder:

codex plugin marketplace add /absolute/path/to/remotion-motionpngtuber

Add this repository as a Claude Code plugin marketplace:

claude plugin marketplace add tegnike/remotion-motionpngtuber

Then install the plugin:

claude plugin install remotion-motionpngtuber@remotion-motionpngtuber

For local development from this folder:

claude --plugin-dir /absolute/path/to/remotion-motionpngtuber

Invoke the skill explicitly in Codex:

$remotion-motionpngtuber Add this MotionPNGTuber character to my video using HyperFrames, the provided asset directory, and VOICEVOX engine.

or:

$remotion-motionpngtuber Add this MotionPNGTuber character to my Remotion composition using the provided asset directory and VOICEVOX engine.

Invoke the plugin skill explicitly in Claude Code:

/remotion-motionpngtuber:remotion-motionpngtuber Add this MotionPNGTuber character to my video using Remotion or HyperFrames, the provided asset directory, and VOICEVOX engine.

The request should include the TTS engine URL, speaker/model/style selection, dialogue lines, and preferred runtime when it matters. If no runtime is specified, choose the one that matches the target project already present in the workspace. If no MotionPNGTuber asset directory is provided, the skill uses the bundled default model.