I've created a simple software to perform similar, using LLM for analysis (finding who is speaking, deciding the voices, translating if necessary) and Gemini for the audio (it is impressive they have good quality no matter the language for the same voice)
What is cool is to then make an AI Skill of it. I've done that and I now can ask on my phone to generate a podcast, or a reading of a page, or small audiobook (very long ones takes few hours) that is sent back to me and I can listen while doing another activity
What is cool is to then make an AI Skill of it. I've done that and I now can ask on my phone to generate a podcast, or a reading of a page, or small audiobook (very long ones takes few hours) that is sent back to me and I can listen while doing another activity