I have been taking this for a spin and it is a phenomenal addition for commutes. One thing I really appreciated was how the ElevenLabs voices tweak their accent when a foreign-origin phrase or foreign name appears in the text and how they modulate the emotion specifically for, say, reporting sad news items. The Listen feature is already at 95/100, but here are a few nitpicky, minor suggestions:
- the “next” button works for advancing to the next story but the compact version of the audio toolbar had the ±10 seconds buttons but no next or previous buttons. I did notice that in the full toolbar mode where all the voice options and other optionalities become visible there is a next story button. Conveniently, the next button on my car works perfectly for going to the next story.
- a headlines-only mode: for feeds like Techmeme which repeat the same text in headlines and then the body.
- offline mode - maybe this will have to be local-models based?
- the listen button should be at the top not the bottom
- I have not tested this out fully and so I may have this wrong: right now, there isn’t a way to cycle through all the posts in a folder of feeds, and that this can only be done on an individual feed basis. It would great to go listen to all the posts at a folder level.