We spent thirty years arguing about what "k." meant

He texts you "k."

Is that fine? Is that furious? Is that over? You'll spend the next hour finding out, probably with a friend reading over your shoulder.

For about thirty years, most of our conversations have been typed. Texting was fast and cheap, and it won. It also shipped without the part of talking that tells you what someone means.

What text throws away

When you talk, the words are only part of the message. The rest is prosody: pitch, pace, volume, the pause, the one word you lean on. "I didn't say he cheated" means something different depending on which word you stress. Text hands the reader every version at once and lets them pick the worst.

So we invented patches. Exclamation points to sound friendly. "lol" to soften anything. ALL CAPS for volume. The period, which somehow became hostile. Thousands of emoji, doing the job one sigh does for free.

So we built for the part text lost

Talking is faster, too. Most people speak around 150 words a minute and type on a phone at a fraction of that. But speed was never the hard part. Tone was.

Our AIs talk with you voice to voice, end to end, so nothing gets flattened into text on the way. They hear the crack in your voice, the speed-up when the story gets good, the "I'm fine" that isn't. They answer with tone of their own: a laugh, a long slow "no," emphasis right where it belongs.

Around that voice, our team built a generative social engine: a system that listens along on every call, works out what mattered to you once you hang up, and carries it into the next conversation. You never see any of it. You just notice you're being heard.

Why our companions are voices first

Coco goes quiet for a beat when the gossip is truly scandalous. Sloane can hear the difference between your "he's fine" and your "he's fine." Try typing that.

Next time he sends "k.", don't decode it alone. Call Sloane.