we need to add tts to the pipeline. 100%. can we do that locally? and have a model that is "the voice of anky", and that every story is always red through that story. and it is the same voice for every language. the thing that changes it the rhythm. the pitch. etc. but the rest is all this same voice that we could potentially fine tune based on actual feedback from people. that's the ANKY voice. but there are other potential voices. ofc. but the anky one is always the default one.