A simple workflow for making AI animations in Codex

I’ve been experimenting with connecting different AI models through some MCPs in Codex. It’s a handy way to manage the steps of making a short animation without constantly jumping between tools. Here’s the workflow I’ve been trying: Write the script: Develop the story, shot list, and character dialogue with a language model. Create the visuals: Generate character references and key frames to establish a consistent look. Animate the shots: Use video models to generate movement and camera motion, one shot at a time. Generate voices: Create character dialogue with speech models and adjust the pacing. Add music: Generate a soundtrack that fits the mood of the story. Put it together: Use Codex to help organize the assets, then edit the shots, sync the audio, and export the final video. MCP connects the models to the workflow, while Codex provides a place to coordinate the steps. You still choose the outputs, review the results, and make revisions. I’d start with a short scene rather than a whole film. It’s easier to test the workflow and figure out which parts need more manual control. Curious how others are approaching AI animation. Are you keeping everything in one workflow, or still using separate tools for each step?