This is a new and improved version of my previous script: https://ecchi.iwara.tv/forums/script-unpack-vmd-txt But I think it is different enough to deserve a separate post.
The biggest difference is that it now can convert both directions!!! So you can convert a motion to TXT, then edit it as you wish, then convert it back to VMD and load it into MMD and actually run it. Very cool!
Also, I now support all frame types in the VMD format: header data (version and model name), boneframe data (bone position, angle, physics on/off, interpolation curve, etc), morphframe data (morph name and value), cameraframe data (camera position, angle, FOV, interpolation curve, etc), lightframe data (light source color and direction), shadowframe data (global setting for self-shadow mode and shadow range), ikdispframe data (IK bone enable/disable and model hide/display)
Also also, I added a README file that explains a bunch of stuff like how to run it and what you can do with it.
Still requires you to have Python installed of course