Feature overview
| Feature | Availability | What to know |
|---|---|---|
| Online real-time transcription | 6 languages | Displays text as you record and requires an internet connection. |
| On-device real-time transcription | Varies by Mac | HyNote reads the on-device speech languages currently provided by macOS. Apple currently lists 10 languages for transcription in Notes as a public reference range. |
| Online post-recording transcription | 36 languages | Six also support online real-time transcription; the remaining languages are processed after recording ends. |
| Local AI | Apple currently lists 16 | HyNote uses the system Local AI capabilities available on the current Mac. |
| App interface | 9 languages | The number of interface languages is different from the number of content languages Local AI can process. |
| Online real-time speaker labels | Supported | Distinguishes speakers in real time but does not identify their real-world identities. |
| Local speaker identification Beta | Supported | Runs after an on-device transcription recording ends and, with your confirmation, can suggest matches to saved speaker names in future recordings. |
Transcription languages
Online real-time transcription
6 languagesOnline real-time transcription displays text while you record and requires an internet connection. It is currently available on Plus and Unlimited plans.
With speaker labels enabled, HyNote can also distinguish different speakers in the live transcript.
If you select another online language, HyNote will not show live text, but it will still save the recording and process it after recording ends. If your connection drops, audio already saved on your Mac is not deleted.
- English
- Spanish
- French
- German
- Italian
- Portuguese
Online post-recording transcription
36 languagesThe 36-language count describes the recording language options and the range available for post-recording transcription. It does not mean that all 36 languages can display text in real time while you record.
- Arabic
- Chinese
- Croatian
- Czech
- Danish
- Dutch
- English
- Persian
- Finnish
- French
- Galician
- German
- Greek
- Hebrew
- Indonesian
- Italian
- Japanese
- Korean
- Latvian
- Malay
- Norwegian
- Polish
- Portuguese
- Romanian
- Russian
- Serbian
- Slovak
- Somali
- Spanish
- Swedish
- Tagalog
- Thai
- Turkish
- Ukrainian
- Vietnamese
- Zulu
On-device real-time transcription
Varies by MacHyNote does not lock on-device real-time transcription to a fixed number of languages. It asks macOS which languages are currently available for on-device processing on this Mac and shows only the options returned by the system.
Apple currently lists the following 10 languages for audio transcription in Notes. This is a public reference range for Apple's feature, not a promise that every Mac running HyNote will show exactly these 10 languages or that HyNote is limited to them.
- English
- Spanish
- Portuguese
- Italian
- French
- German
- Japanese
- Korean
- Simplified Chinese
- Traditional Chinese
System requirements
- Minimum system
- macOS 14 or later.
- Mac requirement
- The current Mac must actually provide on-device speech recognition for the selected language.
- macOS 14 or 15
- You can use a language when the system already provides on-device speech recognition for it. Available languages can differ between Macs.
- macOS 26 or later
- Supports the newer downloadable language models. The first use of a language may require an internet connection for the download; after that, processing can run on device.
- Language and region
- Device language, region, and installed system speech resources can affect the available options. Apple's 10-language description for Notes applies to Macs with M1 or later whose device language is set to one of those languages; it does not define a fixed HyNote limit.
- Permissions
- Allow HyNote to use the microphone and, when requested by the system, speech recognition.
A webpage cannot read your Mac model, macOS version, installed language models, or permission state. The on-device language list shown in HyNote is therefore the final indication of what is available on that Mac.
Read Apple's Mac transcription requirementsLocal AI languages
Apple's current language range
16-language referenceHyNote does not maintain a Local AI language list independent of macOS. It uses the system Local AI capabilities available on the current Mac. Apple currently publishes the following 16-language range.
When the system determines that the input and output languages are supported, Local AI can be used for writing, summarization, chat, title generation, and summary generation. You can also request an answer in another supported language in your prompt.
Output usually follows the note content, your prompt, or the task instruction. English, Spanish, French, Portuguese, and some other languages may include regional variants; those variants are not counted more than once.
HyNote checks whether the system model and current app language are available. The system may refuse a request when the input or requested output language is unsupported. Apple's list is a reference range, not an unconditional guarantee for every Mac, feature, or mixed-language input.
- English
- Danish
- Dutch
- French
- German
- Italian
- Norwegian
- Portuguese
- Spanish
- Swedish
- Turkish
- Simplified Chinese
- Traditional Chinese
- Japanese
- Korean
- Vietnamese
App interface languages
9 languagesThe app interface language and the content languages available to Local AI are separate capabilities. If the interface is not available in a language, that does not mean Local AI cannot understand it.
HyNote does not currently have a separate AI output language setting. To request a specific output language, include it directly in your prompt.
- English
- Japanese
- Spanish
- French
- German
- Italian
- Korean
- Dutch
- Simplified Chinese
Requirements and regional availability
To use Local AI in HyNote, all of the following conditions must be met:
- An Apple silicon Mac
- macOS 26.4 or later
- Apple Intelligence is turned on
- The system models have been downloaded and are ready
- The device language and Siri language are set to the same supported language
- The current app language is supported
- The current device and region meet Apple's availability conditions
According to Apple's current guidance, Apple Intelligence is not currently available on supported devices purchased in mainland China. For supported devices purchased outside mainland China, it is also not currently available when the user is in mainland China and the Apple Account country or region is set to mainland China.
Check Apple's language and regional availabilitySpeaker features
Online real-time speaker labels
Real timeOnline real-time speaker labels work with online real-time transcription in its six supported languages.
The feature marks different voices as Speaker A, Speaker B, and so on. It answers which words came from the same speaker, but it does not determine a speaker's real name or automatically associate live labels with speaker profiles saved on your Mac.
Local speaker identification Beta
On device · BetaLocal speaker identification runs after a recording made with on-device transcription ends. It does not run as a post-recording local identification step when Online Transcription is selected; online real-time speaker labels are a separate feature.
The feature analyzes voice characteristics rather than transcript language, so it does not have a fixed language list. It requires an Apple silicon Mac, Local Speaker Identification Beta enabled, and a one-time download of an approximately 22 MB local model.
Speaker labels still need to align with the on-device transcript. Whether the words display correctly therefore depends on whether the current Mac supports the recording language.
- Distinguish different speakers in a recording
- Listen to samples and name speakers manually
- Adjust, merge, or ignore speakers
- After you confirm a speaker and choose to remember them, suggest a matching saved name in future recordings
Local speaker identification is an assistive labeling feature, not identity authentication. Overlapping speech, short responses, background television, echo, distant recording, similar voices, or a microphone change can affect separation and name matching. Review important content yourself. Do not use this feature for access control, payments, evidence, or any other scenario that requires reliable identity authentication.
Privacy and processing
- Online real-time and post-recording transcription
- Audio must be sent to an online transcription service for processing.
- On-device real-time transcription
- Transcription uses the on-device speech capabilities provided by macOS.
- Local AI
- Prompts and model generation are processed on a compatible Mac.
- Local speaker identification
- After the model download is complete, analysis runs on your Mac. Speaker profiles you choose to remember are stored on that Mac.
On-device transcription and Local AI describe individual processing steps; they do not change where a note is saved. If you choose a cloud note, its recording, note content, and saved AI results can still upload or sync through the cloud workflow. Choose On This Mac before you start when the entire note must remain on the current Mac until you explicitly add it to HyNote.
How to read the language counts
- Regional variants of English, French, Spanish, and Portuguese are not counted more than once.
- A supported language is one the feature can process. It does not mean every language, accent, recording environment, and task will have exactly the same accuracy.
- On-device capabilities supplied by macOS can change with the system version, device, region, and model updates. HyNote uses the availability currently reported by your Mac.