Lightricks/LTX-Video ? reverse-engineered prompt

I want to build a tool for creating AI videos. The main thing is that it should be able to generate video from just a text prompt, but also let me start with an image and animate it. It's really important that it can generate synchronized audio along with the video, so the sound and motion match up perfectly.

The quality needs to be high—think 4K resolution and high frame rates for smooth motion. I also want a lot of creative control. Let me guide the generation by setting multiple keyframes, or even use things like depth maps or character poses to direct the animation. It should be set up so I can run generations locally using a Python script. Please make it easy to switch between different models, like a fast one for previews and a more powerful one for the final output. You can look up the latest docs for any dependencies.