It’s too complicated to explain exactly on a phone keyboard but it’d be pretty fuckin easy.

Step 1:

(Option1) Record a video of a woman with a similar body type saying whatever you want

(Option 2) generate a video of the above. Not preferable since 10 seconds is kind of the max length reasonable

Step 2:

Take a screenshot of the AI girl and run a face replacement.

Step 3:

Do a voice clone.

You can learn to do it with comfyui in less than a day if you have the GPU for it.