Nasu - Private, On-Device Transcription

Nasu turns your conversations, sessions, and voice notes into clean, searchable text — and everything happens on your device. No cloud processing, no account, no subscription servers. Nothing you record ever leaves your hands.

Built for people whose recordings are genuinely confidential: therapists and counselors, physicians, lawyers, researchers, journalists — and anyone who thinks a voice note shouldn't be someone else's training data.

Key Features

• Private by architecture: transcription, speaker identification, and summaries all run entirely on your device. Works fully offline — airplane mode is a feature, not a limitation.
• Who said what: Nasu tells voices apart and labels every turn in the transcript. Name a voice once and it's recognized in future recordings; tap any label to correct it.
• Voice profiles you control: saved only with consent, end-to-end encrypted, and deletable any time.
• Written up, instantly: short summaries, titles, and tag suggestions — generated on your device with Apple Intelligence (on supported devices).
• Clean transcripts: filler words ("um", "uh") removed automatically — reversibly. Custom vocabulary for the names and terms that matter.
• A library that finds things: folders, tags, and full-text search across every word you ever recorded.
• Your notes, your format: export formatted PDF or Markdown, or copy plain text. Share audio in from any app.
• All your devices: native on iPhone and iPad (side-by-side library and transcript view), runs on Apple silicon Macs.
• Local or iCloud storage: local by default; iCloud sync is strictly opt-in for synchronization across your devices.
• Languages: English, French, German, Italian, Portuguese, and Spanish — carefully validated rather than promised — with automatic language detection.

Versions

Version 4.1 — Nasu knows who said what

• Speaker labels: Nasu tells voices apart and labels every turn in the transcript — entirely on your device
• Name a voice once and Nasu recognizes that person in future recordings; it gets better with every correction
• Voice profiles are consent-gated, end-to-end encrypted, and deletable in one swipe
• A transcript editor: fix any wording, reassign any speaker, replay any line with a tap
• On-device summaries, smarter titles, and tag suggestions with Apple Intelligence (on supported devices)
• "Two-Person Sessions" mode for the most accurate results in 1:1 conversations

Version 4.0 — A ground-up redesign

• A new calm interface, in light and dark, built for focus
• Native iPad support with a side-by-side library and transcript view
• Folders, tags, and full-text search across all your transcripts
• Custom vocabulary, formatted PDF/Markdown export, automatic filler-word removal
• Privacy, tightened: iCloud sync is now strictly opt-in (off by default), and analytics can be disabled

Version 3.0 — AI-powered title generation

• Automatically generated titles, powered by AI entirely on your device
• Improved interface and feedback when switching storage between iCloud and local