Does real-time Japanese translation work well?

Yes. AddSub uses on-device speech recognition tuned for Japanese and Apple's Neural Engine translation. For clear speech — news, anime, lectures, podcasts — English subtitles appear in under a second.

Does it work with anime?

Generally yes. AddSub captions the audio track, so it works with Crunchyroll, Netflix, and local anime files. Heavy background music or very fast speech can reduce accuracy.

Can I see the original Japanese and the translation together?

Yes. The overlay shows bilingual captions — original Japanese plus the English translation — with speaker labels.

Is my audio sent to a server for translation?

No. Everything runs 100% on-device. No Japanese audio is ever uploaded.

Japanese → English real-time translation on macOS

The short answer: AddSub gives you instant English subtitles for any Japanese content playing on your Mac — anime, news, lectures, live streams, podcasts — translated live, 100% on-device. Set the pair once (日本語 → English) and read everything as it’s said.

Why Japanese → English is AddSub’s flagship pair

Japanese media has a huge global audience — anime, drama, J-pop interviews, and news — and much of it lacks good real-time English subtitles. AddSub was built around this: its on-device speech recognition is tuned for Japanese, and translation runs on the Neural Engine, so you get natural English captions as the speaker talks. No waiting for fansubs, no missing a beat in live broadcasts.

How it works, step by step

Install AddSub (download) — free 7-day trial, no payment to start.
Set the language pair — Settings → Languages → source: Japanese (or Auto-Detect), target: English. Bilingual captions show both lines.
Play anything Japanese — Crunchyroll, Netflix, YouTube, local anime files, live streams. English subtitles float above the video with speaker labels.

What it handles well

News & documentaries: excellent — clear narration, formal register.
Anime: very good — captions the audio track; works on Crunchyroll/Netflix/local files.
Lectures & podcasts: excellent.
Live broadcasts & interviews: good — includes speaker labels (S1, S2, S3…).

Note on accuracy: clear speech produces the best results. Dense background music, overlapping voices, or heavy slang can lower accuracy — a limitation of all on-device speech recognition today.

Privacy & cost

Translation happens entirely on your Mac — no Japanese audio is uploaded, there’s no account, and there are no per-minute fees. After the 7-day trial, a one-time license keeps it yours.

Read every Japanese video — free for 7 days

No payment needed to start. Download for macOS.