Hi Product Hunt. I am Andrey, I built Kekoso on my own.
I write in two languages all day and I kept hitting the same wall: the dictation
that is good is in the cloud, and the dictation that is local is one engine with
one set of languages. So I built the thing I wanted. Kekoso runs three engine
families side by side. Parakeet TDT v3 is the fast one and covers 25 European
languages. Whisper through WhisperKit is the accurate one and covers the long
tail of 99. SenseVoice handles Chinese, Cantonese, Japanese and Korean, which is
exactly the gap Parakeet leaves. You download a model once, switch whenever, and
every job after that uses whichever one you picked.
The part that took the longest was silence. Whisper was trained on subtitled
video, so on a pause it writes what subtitles usually say there: sign-off lines,
translator credits, "to be continued". Pasting that into a message you send a
second later is how you lose trust in a dictation app. Lines with no audio behind
them now get dropped before anything reaches your text field.
Everything runs on your Mac. The app talks to the network in three places only:
your licence, app updates, and fetching audio when you paste a YouTube link.
Turn Wi-Fi off and dictation, files and call recordings all still work. No
account anywhere.
Being honest about the limits: Apple silicon and macOS 14.4 or later, no Intel.
It is not in the App Store, you download a notarised build from the site. There
is no speaker diarization yet, so it separates nothing in a two-person recording.
$29 once for 3 Macs, 7 days to try it, no subscription.
Welcome to StartupBase
Join a community of founders sharing their products and discovering what others
are building.
Email sign-in is for existing accounts. New here? Use Google, LinkedIn, or X above
— your account is created automatically.
Continue with email
Enter the email address on your existing account and we'll send you a one-time
sign-in code. This does not create an account — to sign up, go back and
continue with Google, LinkedIn, or X.
Enter the verification code we sent to your email address.
Continue with Google
Continue with LinkedIn
Continue with X
Hi Product Hunt. I am Andrey, I built Kekoso on my own.
I write in two languages all day and I kept hitting the same wall: the dictation that is good is in the cloud, and the dictation that is local is one engine with one set of languages. So I built the thing I wanted. Kekoso runs three engine families side by side. Parakeet TDT v3 is the fast one and covers 25 European languages. Whisper through WhisperKit is the accurate one and covers the long tail of 99. SenseVoice handles Chinese, Cantonese, Japanese and Korean, which is exactly the gap Parakeet leaves. You download a model once, switch whenever, and every job after that uses whichever one you picked.
The part that took the longest was silence. Whisper was trained on subtitled video, so on a pause it writes what subtitles usually say there: sign-off lines, translator credits, "to be continued". Pasting that into a message you send a second later is how you lose trust in a dictation app. Lines with no audio behind them now get dropped before anything reaches your text field.
Everything runs on your Mac. The app talks to the network in three places only: your licence, app updates, and fetching audio when you paste a YouTube link. Turn Wi-Fi off and dictation, files and call recordings all still work. No account anywhere.
Being honest about the limits: Apple silicon and macOS 14.4 or later, no Intel. It is not in the App Store, you download a notarised build from the site. There is no speaker diarization yet, so it separates nothing in a two-person recording.
$29 once for 3 Macs, 7 days to try it, no subscription.