More accurate transcriptions: the language is detected once per audio and no longer switches mid-message, and the high-accuracy model is used automatically once it is ready on your iPhone. More useful summaries that understand who is speaking, what is being asked and how the audio ends, with a length that fits its duration. Fixed split numbers such as 20,000 and incomplete or repeated key points.

Record from Dynamic Island, the Lock Screen and widgets. Improved transcription and summaries with global context and local personal memory. Better recording recovery and interruption handling. New optional usage metrics and crash diagnostics preferences, without sharing your audio content.

Audio Copilot now requires paid access: • Full Pro access: recordings up to 4 hours, transcripts, ideas, lists, and tasks. • A 7-day free trial on the annual plan for eligible users. No ongoing free plan. • Existing purchases and saved data are retained. Restore purchases and view or export your saved content. • Reliability improvements for recording, sharing audio, and background processing.

Audio Copilot 2.0, redesigned from the ground up: • Record meetings and classes up to 4 hours long with your iPhone locked, plus live transcription (iOS 26) and the Dynamic Island. • Tells speakers apart and gives you a summary, decisions, and action items with owners. • Speak: share an idea, say “add milk to my shopping list” or “remind me to call Josh tomorrow at 10,” and everything goes where it belongs. Ideas on the same topic group together on their own. • Lists and tasks with reminders, interactive widgets, Control Center, the Action Button, and Siri. • Your WhatsApp voice messages with what's being asked of you and a reply ready to go. • Face ID lock and search across your audio, ideas, and lists. • More for free: unlimited Speak and voice notes up to 5 minutes, and free meetings up to 30 minutes. Everything still runs on your iPhone, with no cloud.

Minor improvements.

Audio Copilot 1.1: meetings with automatic notes and a more reliable on-device engine. • Record meetings and conversations from the app. When you stop, you get a summary, decisions, action items with owners, open questions and a ready-to-send follow-up message. • Mark key moments while recording and jump to them from the notes. • Recording is fail-safe: if the app closes or a call comes in, your audio is kept and recovered. • Export meeting notes as Markdown or share them with one tap. • New onboarding with pictures that shows step by step how to share audio from WhatsApp. • Processing no longer gets stuck: it watches its own progress, retries on its own and tells you clearly when an audio has no speech. • Meeting notes are free up to 10 minutes; unlimited with Pro.

More accurate on-device transcription, plus several important fixes. • iPhones with 8 GB of memory or more now use the Large v3 Turbo local model, which is far more accurate on difficult audio and across languages. If you already had the previous model, you can switch any time from Settings › On-device AI. • Fixed: 16 kHz mono AAC and Opus voice notes are now converted properly before transcription. • Fixed: Chinese, Japanese, Korean, Arabic and Russian keep per-sentence playback timings. • Fixed: automatic language mode no longer falls back to Spanish when retrying. • Your imported audio is no longer deleted automatically after seven days. • Lower memory use on long recordings, and reliable summaries for long transcripts. • Fixed: the daily limit prompt can now be dismissed and no longer reopens by itself. • Fixed: plan prices are now readable in both light and dark mode. • Transcription errors now explain what went wrong and let you remove the affected audio.

Maintenance update with safer Pro entitlement validation, clearer StoreKit catalog handling, and purchase and restore telemetry improvements.

Privacy improvements: opt-in anonymous telemetry (TelemetryDeck). No tracking, no IDFA.

Stability improvements when opening audio from messaging apps, widgets, links, and the share extension. Supported languages, iOS integration, and approved Pro subscriptions have also been tightened.

Improved stability when importing from messaging apps, smoother karaoke playback, a cleaner Pro flow with monthly and annual subscriptions, and quality refinements for transcription, summaries, and key points.

• Simplified onboarding with one balanced model • Clearer library with read/unread state • More careful summaries and key points grounded in the audio • Better source context when sharing audio from other apps

• New in-app audio summary • Improved transcription accuracy • Better WhatsApp voice memo support