The 30 transcription languages
Every language below gets live captions while recording and can be recognised again after the session on Plus and Pro.
| Language | Interpretation from it | Interpretation into it |
|---|---|---|
| Arabic | Yes | Yes |
| Cantonese | Yes | No |
| Czech | Yes | Yes |
| Danish | Yes | No |
| Dutch | Yes | Yes |
| English | Yes | Yes |
| Filipino | Yes | Yes |
| Finnish | Yes | No |
| French | Yes | Yes |
| German | Yes | Yes |
| Greek | Yes | No |
| Hindi | Yes | Yes |
| Hungarian | Yes | No |
| Indonesian | Yes | Yes |
| Italian | Yes | Yes |
| Japanese | Yes | Yes |
| Korean | Yes | Yes |
| Macedonian | Yes | No |
| Malay | Yes | Yes |
| Mandarin Chinese (Simplified) | Yes | Yes |
| Persian | Yes | Yes |
| Polish | Yes | Yes |
| Portuguese | Yes | Yes |
| Romanian | Yes | No |
| Russian | Yes | Yes |
| Spanish | Yes | Yes |
| Swedish | Yes | No |
| Thai | Yes | Yes |
| Turkish | Yes | Yes |
| Vietnamese | Yes | Yes |
Translation only (11)
Live Interpretation can write these, but MemoFlow cannot transcribe them: Bengali, Burmese, Gujarati, Hebrew, Khmer, Marathi, Tamil, Telugu, Traditional Chinese, Ukrainian, Urdu.
Four lists that answer different questions
“Supports Japanese” can mean four different things in an app like this, and MemoFlow's answer is not the same for each. The table below covers the first three; the fourth is here so it is not mistaken for them.
- Transcription — 30 languages. You can choose them as the Speech Recognition Language for a recording, see live captions in them, and have the session recognised again after it ends.
- Interpretation source — the same 30. Live Interpretation can only translate speech it can transcribe as it happens.
- Interpretation target — 33. The translator writes some languages it cannot listen to, such as Ukrainian, Hebrew and Bengali, and cannot write some it can hear, such as Cantonese, Swedish and Danish.
- Interface — the app is translated into 29 languages and this website into 9. Neither list says anything about transcription.
Choose the language before you record
There is no automatic language detection. The Speech Recognition Language in Settings is the default for AI Meeting recordings, and a recording can use a different one without changing the default. Live Interpretation has its own Translation Source Language and Translation Target Language.
A session listens for one language. In a meeting that switches between two, pick the language most of it will be in; passages in the other language will need checking against the audio.
How this list was checked
The lists on this page are generated from the same language registry the app and servers use, so they change when the product changes rather than when someone edits the page.
After the 12 September 2026 deployment, each of the 30 transcription languages was accepted by a live production session and each of the 33 translation targets returned text. That check shows every language works end to end. It does not measure how accurate any of them is, and MemoFlow has not published accuracy figures for any language.
Limits to know
- The language is chosen before recording; there is no automatic language detection.
- Live captions need a connection. Offline recording continues without them.
- No accuracy figure is published for any language.
- Machine translation, not a human interpreter.
- Recordings in a language outside the 30 are not offered in the AI Meeting picker.
- Language notes exist for Japanese, Chinese, Korean and Spanish so far; the other languages work the same way but have no page of their own yet.