← Blog

Junco's Voices Just Moved Onto Your iPhone

If you use Junco, you might notice something different about your episodes soon. The voice reading your morning briefing is still the same friendly narrator. It's just coming from a different place now.

What changed

Up until now, when Junco turned your newsletters into an episode, the audio itself was generated somewhere out in the cloud, then sent down to your phone. That worked fine, but it meant a bit of waiting while the audio rendered, and it meant your phone needed a connection to get a fresh episode ready.

We just moved that whole step onto your iPhone, using Kokoro, an open-weight on-device text-to-speech model. The premium voices now render locally, right on your device, using the newsletter script Junco already prepared for you. No round trip to a server required.

Why it's better

A few things get noticeably nicer:

  • Episodes are ready faster. Since there's no waiting on a remote server to render audio, your episode can start playing the moment its script is done.
  • It works offline. Once an episode's script is on your phone, Junco can narrate it even if you don't have a signal.
  • It's a step toward more private by design. The audio itself is created right on your device instead of passing through another service along the way.

One small side effect

Because this is a genuinely new narration engine, your default voice got reset as part of the switch. If your favorite voice suddenly changed, that's why, not a bug. Open Settings in the app and you can preview the full lineup and pick whichever one sounds best to you. It only takes a few seconds.

If anything ever goes wrong with the on-device voices (spotty storage, first-time setup still finishing, that kind of thing), Junco will quietly fall back to your iPhone's built-in voice so your episode still plays. You won't get stuck waiting.

The rest stays the same

Your newsletters still arrive the same way, get summarized the same way, and show up in the same queue every morning. This change is about how the audio gets made, not what you get. It should just feel a little snappier.

As always, if something feels off, let us know. We're still a small team shipping fast, and feedback shapes what we build next.