Contact Support
For questions, bug reports, or feature requests, contact:
Support is available in English and Japanese.
For bug reports, include your macOS version, KoeSub version, selected audio source, and the steps that did not work as expected.
Help, troubleshooting, and frequently asked questions
For questions, bug reports, or feature requests, contact:
Support is available in English and Japanese.
For bug reports, include your macOS version, KoeSub version, selected audio source, and the steps that did not work as expected.
Watch the Japanese tutorial, then use the permission checklist below if subtitles do not appear.
KoeSub may request macOS permissions for Screen Recording, Microphone, and Speech Recognition depending on the selected audio source and system configuration.
If translation does not start, open macOS System Settings and confirm that KoeSub has the required permissions.
No. KoeSub may launch at login if enabled, but it stays idle. Audio capture and translation start only when you turn translation on.
Version 2.2 improves translation accuracy. It remains an aid to understanding live conversations, not a guarantee of verbatim or professional interpretation.
No. KoeSub uses Apple system speech recognition and translation frameworks. Results can vary depending on audio quality, accent, speaker speed, background noise, and supported language availability.
Real-time speech recognition updates partial text while someone is still speaking. KoeSub translates finalized segments, so translated subtitles may appear slightly after the spoken words and may be revised as recognition stabilizes.
Yes, when the selected macOS version, device configuration, and permissions support the required audio capture path. Audio quality can vary depending on speaker volume, microphone placement, headphones, and background noise.
No. KoeSub does not upload your audio, transcripts, or translations to Toybird Labs servers.
KoeSub provides five minutes of active subtitle time per launch. Time is counted only while subtitles are running. Quitting and relaunching resets the five-minute allowance.
Custom acoustic ML can recognize evidence of Singapore English and, when supported, adapt speech recognition to en-SG. General English is the default.
No. KoeSub focuses on supporting understanding during live conversations.
KoeSub v2.2.0 supports English-to-Japanese subtitles. Availability depends on Apple system speech recognition and translation resources.
App name: KoeSub
Current release: v2.2.0 · macOS 15+
Bundle ID: com.toybird.KoeSub
The installed app version can be checked in Finder or macOS app information.