Record, with as much on screen as you want.
Choose a simple recording view, a live transcript, or a transcript with Live Recap. On Mac, you can also capture system audio together with your microphone.
ClariaMac iPad iPhone
From a conversation
to a page of notes.
Record a meeting, read along with the transcript, and keep the key points in a short recap. Come back later to edit, listen again, or share your notes.
App Store release in preparation.
During and after the conversation
Use the recording
the way you need.
Already have the audio? Bring it in.
Open an audio or video file, drop it into Claria, or send it from another app using Share. Transcribe a saved recording whenever you’re ready.
Keep the original. Share a clearer version.
Tidy up wording, shorten a passage, or pull out the main points. Switch back to the original transcript to check what was said before copying or sharing.
Mac, iPad and iPhone
The same work, with room to suit your screen.
Use the wider workspace on Mac and iPad. On iPhone, switch between the transcript and recap while keeping the recording controls close.
Actual app screens with fictional sample notes. Select an image to see it at full size.
Keep working after you press Stop
A recording you can find, follow and share.
Live transcript and recap
Read the words as they are recognized, with a recap that follows the conversation’s original language. Scroll back to review, then return to the latest content when you’re ready.
Bilingual captions
Keep the original transcript and add a translation. Choose Traditional Chinese, Simplified Chinese or English in the caption panel. Captions use your selected recap model.
Playback that follows the text
Play a saved recording and follow the matching transcript segment. Following is on by default; turn it off on screen when you want to read at your own pace.
Folders for your meetings
Name folders for a project, class or recurring meeting. Drag recordings to change their order, move them into a folder, or bring them back out. Deleted meetings go to Recently Deleted.
Choose what you share
Share the original transcript, the polished version or the original audio. The source stays available when you need to check a name, a number or the exact wording.
Your own dictionary
Save names, specialist terms and preferred spellings for text polishing. Your dictionary is included with the library’s iCloud sync and backups.
iCloud backup across devices
With iCloud Drive available, backup is on by default. Completed recordings, notes, folders and dictionary changes can travel between your Mac, iPad and iPhone. Check transfer progress in Settings.
Controls that fit the device
On Mac, use the menu bar to check recording status and reach quick controls. On iPad, work beside another app. Microphone recording can continue in the background on iPhone and iPad; imported-file transcription pauses and resumes when you return.
On your device
Speech and text have different jobs.
The speech model turns audio into a transcript. The text model makes recaps, translates captions and polishes wording. Choose speech recognition, recap and text polishing separately in Settings.
| Model | Language support | Availability in Claria |
|---|---|---|
| Apple Speech | Languages available through Apple’s on-device speech service. | Uses system-managed speech resources; availability depends on the device and language. |
| Qwen3-ASR 0.6B | 30 languages, including Chinese, English, Japanese and Korean; automatic language detection. | A smaller local speech model, downloaded in Settings. |
| Qwen3-ASR 1.7B | The same 30-language coverage, with support for 22 Chinese dialects across the Qwen3-ASR family. | A larger local option with higher memory requirements. |
| Parakeet TDT 0.6B v3 | 25 European languages, including English. Chinese is not supported. | A multilingual Parakeet download. |
| Parakeet TDT 1.1B | English only. | A separate English speech model; its larger size does not add languages. |
Apple Foundation Models
Use Apple’s on-device text model for recaps, captions and polishing when Apple Intelligence is supported, enabled and ready. This is separate from Apple Speech.
Gemma 4 E2B · E4B · 12B
Three downloadable text-model options for recaps, captions and polishing. You can use different models for Live Recap and your final polished text. Larger downloads require more storage and memory.
Qwen, Parakeet and Gemma are integrated into Claria and downloaded separately when needed. The App Store version uses Apple-hosted Background Assets for these downloads. Settings shows supported languages, download progress and device restrictions; a model must be installed before it can be selected. Not every model is available on every device.
Language references: Qwen3-ASR · Parakeet 0.6B v3 · Parakeet 1.1B
Your own cloud account
Bring the provider you already use.
Add one or more connections with your own API keys. Use a speech service for transcription and a different text model for recap or polishing, without running those models locally.
| Provider | Speech recognition | Recap, captions & polishing |
|---|---|---|
| OpenAI | Transcription models, including Whisper | Compatible GPT text models |
| Google Gemini | Gemini models that accept audio | Gemini text generation |
| Anthropic Claude | Not offered through this connection | Claude |
| Groq | Whisper models | Compatible text models hosted by Groq |
| Deepgram | Nova speech models | Not offered through this connection |
| AssemblyAI | Universal transcription models | Not offered through this connection |
| Microsoft Azure | Azure Speech | Compatible Azure OpenAI deployments |
| Alibaba Cloud Qwen | Qwen ASR | Qwen text models |
| DeepSeek | Not offered through this connection | DeepSeek chat models |
| OpenRouter | Not offered through this connection | Compatible text models on OpenRouter |
| Custom compatible API | Compatible audio transcription endpoint | Compatible chat completion endpoint |
Enter the model ID available to your account. These are supported connection types and model families, not a guarantee that every model or API version from a provider will work. Language coverage, access and limits depend on the selected model. Azure Speech and Azure OpenAI may need separate connections and keys.
Cloud processing needs an internet connection and sends the relevant audio or text directly to your chosen provider. API usage is billed by that provider, separately from Claria. Keys stay in the device’s Keychain and are excluded from iCloud and exported backups.
What affects the result
The model you choose makes a difference.
The same recording can produce different transcripts, translations and recaps with different models. A larger model is not automatically better for every language or task.
Start with the spoken language.
Use a speech model that supports the recording’s language. Translating the result later cannot repair words the speech model failed to recognize.
Give it clear audio.
Background noise, overlapping speakers, accents and specialist vocabulary can affect recognition. Missing or mistaken words can then carry into the recap.
Check what matters.
Text models can omit details or change meaning. Review names, numbers, dates and decisions against the original transcript and recording before sharing. Model choice also affects speed, memory use and cloud cost.
A few practical details
Before you start.
Can I use Claria on all three devices?
Claria supports Apple silicon Macs with macOS 26 or later, and iPhone and iPad with iOS or iPadOS 26 or later. Some local models need more memory or Apple Intelligence. Local model availability also depends on available memory and downloaded resources.
Can I keep notes between devices?
iCloud backup is on by default when iCloud Drive is available. Completed recordings, text, meeting organization and your dictionary are included. Backups upload in the background, and records from your other devices can be merged into the library. Transfer progress is shown in Settings. You can turn this off in Settings. Models and API keys are not included.
Does recording need a live transcript?
No. Choose the recording-only mode to keep the screen simple, then open the saved recording and start transcription later.
Which files can I bring in?
Claria accepts common audio and video formats, including WAV, MP3, M4A, AAC, AIFF, CAF, FLAC, MP4, MOV and M4V. You can also share a recording from another app or drag a file into the window.
Can I record audio from another app?
On Mac, enable system audio to capture other applications together with the microphone, including when you use headphones. On iPhone and iPad, start capture through the system broadcast picker, then return to Claria to import the captured audio. Other apps may block capture; this does not guarantee recording of phone calls or protected audio. Claria does not save screen video.
Can I use English and Chinese?
The interface supports English and Traditional Chinese. Qwen speech recognition detects the spoken language, and Chinese transcript output can use Traditional or Simplified characters. Optional bilingual captions have three targets: Traditional Chinese, Simplified Chinese and English. Speech-language support depends on the selected model; Parakeet 0.6B v3 does not recognize Chinese, and Parakeet 1.1B is English only.
Have a question about Claria?
Write to us at benson.chuang@tsugas.com.