Launching today
Caliptic Translate — Real
Understand any podcast, video or meeting — 22 languages
2 followers
Understand any podcast, video or meeting — 22 languages
2 followers
Caliptic Translate does two things at once, in 22 languages. Everything your speakers play — podcasts, videos, news, lectures, meetings — becomes live translated subtitles and speech. And everything you say is translated and delivered to the other side as you speak. On a phone, two people can share one pair of earbuds: one language in each ear. It all runs on your own computer — no credits, no quotas, no per-minute billing. macOS, Windows and Android; iPhone in review.
Hey Product Hunt 👋
Caliptic Translate does two things at once, and the second one is the part most tools skip.
Everything coming out of your speakers gets translated. A podcast everyone in another country is talking about, a lecture from a university that doesn't teach in your language, a news broadcast, a documentary, a YouTube channel nobody subtitles, a call you only half follow — all of it becomes live subtitles and spoken translation as it plays. This is what I use it for most, and honestly it's the use case I underestimated when I started building.
Everything you say gets translated and delivered to the other side. Through a virtual mic, a meeting hears you in their language while you keep speaking yours. Face to face, the person across the table just hears it out loud. On a Mac that outgoing voice can even be a clone of your own, generated on-device with Apple Personal Voice.
On a phone there's one more thing, and it's my favourite. Two people can share a single pair of earbuds. Couple Mode sends one language to the left bud and the other to the right, so you each take one and hear only your own. No passing a phone back and forth, no putting a stranger's voice on speaker in a quiet restaurant, and no $200 translator earbuds — you already own the hardware. The app shows which language is in which ear (TR ◀ ▶ EN) and you can swap sides in a tap. This one is mobile-only by nature: Android today, iPhone in review, no desktop equivalent.
The other half of the phone experience is ambient mode — nobody taps a button to say whose turn it is. It detects which of your two languages was just spoken, sentence by sentence, and translates the other way. That sounds like a small thing until you've tried to hold a real conversation while both people keep reaching for the same phone.
It runs on your own computer. No credits, no quotas, no per-minute billing — the three things that make every cloud translator exhausting to actually live with. That's also why it keeps working on a plane with no signal, why there's no account, and why the free tier can be genuinely free rather than a trial. Nothing you hear or say leaves the machine, which matters in a work meeting — but the reason I built it this way was the metering, not the paranoia.
22 languages, including full right-to-left Arabic. Android, macOS and Windows today; iPhone is in App Store review.
Pricing: the free tier isn't crippled — 40-minute sessions on desktop that you reset in a click, no daily cap, no sign-up. Pro removes the reset and adds the spoken voices plus a higher-accuracy model.
Requirements: macOS 15+ on Apple Silicon, or Windows 10 21H2+ (64-bit with AVX2, 8 GB RAM). Models download once (~1 GB), then it's fully offline.
I'm the solo maker and I'll be here all day. Two things I'd genuinely like to hear: what do you listen to that never gets translated? And how does it handle YOUR language pair — tell me which one and I'll answer honestly, including where it's weak.