From mic, apps & files — into one local notebook

Audio Note supports file transcription, realtime microphone capture, realtime app capture, global realtime subtitles, link transcription, and watch-folder automation. There is also Dictation (Beta): press a shortcut and speech lands in the focused input field.

Dictation uses a shortcut to speak into the focused input field. Realtime transcription is for meetings and captions and produces a notebook. Dictation is available to signed-in users and is labeled Beta at every entry point. See the dictation docs for plan limits.

No. It runs offline on your device after a note is transcribed, and requires Pro or Lifetime.

The default transcription workflow runs locally, so your audio and transcripts stay on your device. Content is only sent out when you intentionally use a cloud AI feature.

Audio Note supports Chinese, English, Japanese, Korean, Russian, and many other languages, with automatic detection when the selected model supports it.See supported languages

The app includes device benchmarking and model recommendations. Older machines can start with lighter models or Sherpa for realtime capture, while faster machines can switch to higher-accuracy options.

GTX 10-series and newer NVIDIA cards can use the CUDA engine for Whisper GPU acceleration. Older cards may crash on Vulkan—the app will suggest switching to CUDA in Settings → Transcription → GPU engine. See the docs page “NVIDIA GPU & CUDA compatibility” for details.

They live under Settings → Models → ML Engine. Standard includes SenseVoice and Moonshine, which work for both realtime capture and file transcription. Pro adds FireRed CTC, Qwen3-ASR, Cohere and Parakeet, aimed at higher-accuracy file transcription.

Backups include notes, transcription text, task history, AI chats, and common settings. They do not include raw recordings, downloaded models, or exported caption files. Create encrypted backups under Settings → Backup & Restore—see the “Backup & Restore” doc for details.

For local audio/video files, link transcription, and realtime transcription, the transcription engine runs on your device. Account, authorization, subscription, payment, and feedback features connect to Audio Note servers when you use them.

Realtime transcription fits meetings, livestreams, and cross-app subtitles. File transcription is better for historical media, higher final accuracy, translation, export, and post-processing.

You can export to SRT, VTT, Markdown, TXT, JSON, and other common formats for subtitles, archives, and follow-up processing.

Yes. Audio Note lets you review text, organize notes, translate content, and send the result into AI summaries, outlines, or follow-up writing.

Link transcription works well for podcasts, courses, and public videos. Watch folders are built for automated workflows where recorders, meeting systems, or download directories keep producing new files.

New accounts get a 7-day free trial so you can evaluate local transcription, realtime capture, and the rest of the workflow.

Audio Note is designed around local-first transcription: imported files, linked media, microphone audio, app audio, and system audio are transcribed on your device. Account, authorization, subscription, payment, and feedback features use Audio Note servers. AI Chat calls a cloud model when you use the official provider, but Audio Note does not collect your chat data.

When you use the official AI Chat provider, requests are processed by a cloud model. Audio Note does not collect your chat data.

You can contact us by email or by submitting feedback:Feedback