Generated and spoken entirely on your iPhone. No internet. No accounts. No ads.
A pocket-sized studio that writes and narrates a fresh episode the moment you tap a topic.
Tap a topic and a roughly thousand-word podcast streams live, word by word, in seconds.
Five neural Piper voices speak the episode in real time as it's written — no robotic delays, no cloud round-trips.
Watch the current sentence light up as it's spoken. Past lines dim. Future lines wait their turn.
Touch a word and the narration restarts from that sentence. Scrub through ideas like a Spotify track.
Reached the end? Continue the same thread — the model picks up with the previous paragraph as context.
0.75×, 1×, 1.25×, 1.5× and 2×. Pick a pace, the voice stays natural.
Five on-device voices, bundled with the app. No downloads, no subscriptions.
A whole studio runs on the phone in your hand.
Tap a featured card, browse Recent, or type your own search. Ten fresh angles appear every time.
A bundled Llama 3.2 1B language model drafts the episode on-device, streaming each word as it arrives.
Sentences are handed to the neural voice engine and spoken aloud in real time. No buffering, no servers.
Nothing you type, hear, or generate ever leaves your iPhone. Every word is written and spoken locally — the app works in airplane mode.