Its just a game engine without any IDE. So the typical workflow would be to build your assets and scene in Blender (or Maya, etc.) and then code up the game mechanics in Python or C++. I am tinkering with it for non-gaming (Reinforcement Learning) use cases and I was curious what game devs think of it.
Now there's a name I haven't heard in a very long time... I'm not in game dev right now, though, so grain of salt and stuff. But I did play around with Panda a long time ago, and thought it was fine for getting some 3D stuff going quickly without dealing with the headaches of OpenGL. Similar to PyGame but for 3D. It's neat that it's still seeing development.
Besides the old Pirates of the Caribbean game, which I can't even find mentioned on their site anymore, have there been any commercial games developed with it? I was always under the impression that it was primarily targeted towards CMU students in their game design curriculum. (DigiPen, a school most known for game programming, has in its main degree program students building their own engines from scratch, but there are at least 3 in-house engines that have been developed over the years meant for freshmen and game design students.)
I tend to treat the "commercial use" barrier as the key one to initially evaluate "niche" engines. Like, Ogre3D is fairly "niche" too, but it's been used in many commercial games. That says nothing directly about the relative quality, but it does imply some good things if people were willing to risk a business product on it.
Personally I'd consider Panda again for a game jam or as you mentioned a non-game-but-3D-needing thing, but where for both I'm also expecting to use Python. But I'd want to check out Python bindings to Ogre and whatever else is out there now these days too. (Godot has been top of my "if I ever want to just make a game and not spend time in the weeds doing things from scratch in a non-mainstream language, use this" tools.)
Its just a game engine without any IDE. So the typical workflow would be to build your assets and scene in Blender (or Maya, etc.) and then code up the game mechanics in Python or C++. I am tinkering with it for non-gaming (Reinforcement Learning) use cases and I was curious what game devs think of it.