New in 1.7.5 Make a voice clone from a file: on the new voice screen, tap Import to use an audio file, or a video from Photos or Files. Only the first minute is used. Delete a voice clone: press and hold it in Voices and choose Delete Voice. Its reference recording is removed too, and projects already made with it keep their finished sound. Videos you change the voice of now stay in sync with the picture. A progress screen covers the app while a file is imported, and a failed import no longer leaves an empty editor behind. Recording now works with more microphones, including Bluetooth headsets. Long text for speech asks before it is cut at 5 minutes, and the app no longer freezes while generating speech when the iPhone is low on memory. If finished results are removed when the app restarts, a notice now tells you. VocalShift now requires iOS 18 or later.

New in 1.7.4 Video projects: a video you change the voice of is now saved as a video. Play it back with the picture, or export it as a video or just the audio. Trims and fades keep the picture in sync. Faster and easier on the battery: Text to Speech, AI voices and your own voice clones now use much less processing power, so results come sooner and your iPhone stays cooler. A new voice clone saves right away, with no download to wait for. Your projects are now included in your iPhone backup, so they come back when you restore or move to a new iPhone. To leave them out, turn on Skip iCloud Backup in Settings.

New in 1.7.3 Play while it converts: your result starts playing within seconds instead of waiting for the whole song. Longer songs now use far less memory. Crowd cheering and backing shouts no longer turn into the character's voice. Fixed a crash on some iPhones when creating an AI character voice. If VocalShift ever stops unexpectedly, email us at support@taxlite.net — we read every message.

New in 1.7.2 Fixed crashes when converting with AI voices. Voice conversion now uses far less memory, so it runs reliably on more devices. Long recordings with your own cloned voice and long songs now process without running out of memory. Still 100% offline.

New in 1.7.1 Fixed a rare crash that could happen during voice conversion on devices running low on memory. Smaller stability improvements under the hood. Still 100% offline.

New in 1.7.0 Keep creating while VocalShift works. Voice conversion, song separation, text generation, and voice cloning now share a background processing queue, so jobs continue when you leave the editor and run safely one at a time. Saved projects now reopen and play more quickly, with clearer progress, cancellation, and queue controls. Also includes smoother playback, navigation, localization, and stability improvements. Still 100% offline.

New in 1.6.0 Voice conversion and full-song vocal separation are now significantly faster, especially when trying different AI voices. Processing now shows clearer progress and can be cancelled, with improved stability and memory use for longer recordings.

New in 1.5.1 Several neural voices have been retrained for more natural, distinctive, and consistent transformations. Waveforms now feature a cleaner bar-style design with smoother playback animation and clearer audio detail.

New in 1.5.0 — Sing Along to Full Songs Import a full song with instrumental music and VocalShift automatically separates the vocals, then converts them into Deep Giant, Sweet Girl, or one of your own cloned voices. Also new: one simpler import flow — record or import, then just pick Clean Vocal or Full Song — plus a smoother, more responsive tab bar. Still 100% offline. No accounts, no ads, no tracking.

New in 1.4.0 — Sing in a Different Voice Record or import a singing clip, then convert it into Deep Giant, Sweet Girl, or one of your own cloned voices. Also new: real portraits for every neural voice, and Library now brings your recordings and cloned voices together in one place. Still 100% offline. No accounts, no ads, no tracking.

New in 1.3.3 — Stability Fixes Fixed a crash that could happen when switching between tabs quickly. Still 100% offline. No accounts, no ads, no tracking.

New in 1.3.2 — Bug Fix Fixed British English (Emma, Daniel) voices in Text to Speech, which could fail to generate audio. Still 100% offline. No accounts, no ads, no tracking.

New in 1.3.1 — Malay & Arabic VocalShift now speaks 12 languages — Malay and Arabic join English, German, French, Korean, Japanese, Spanish, Chinese, Portuguese, Italian, and Russian. Also: general stability improvements, including a fix for a rare crash when switching tabs quickly. Still 100% offline. No accounts, no ads, no tracking.

New in 1.3.0 — Clone Your Own Voice Record a short sample and generate a personal voice you can use anywhere in the app — for yourself, or anyone who's given you permission. No training, no upload, ready in seconds. Also new: 6 realistic AI character voices (Deep Giant, Sweet Girl, Soft Lady, and more) that go beyond pitch and effects to truly sound like someone else, plus a major quality upgrade to neural text-to-speech across 7 languages. VocalShift now speaks 10 languages — Spanish, Chinese, Portuguese, Italian, and Russian join English, German, French, Korean, and Japanese. Still 100% offline. No accounts, no ads, no tracking.

New in 1.2.1 — Crash Fix Fixed a crash that could happen right when you tapped Record. Recording is now stable and reliable. Still 100% offline. No accounts, no ads, no tracking.

New in 1.2.0 — More Convincing AI Voices We rebuilt how our neural voices handle pitch, so characters actually sound like the character: • Deep Giant is now dramatically deeper and more commanding • New character: Sweet Girl — a bright, youthful voice • Narrator retuned for a clearer, more natural read Also in this release: • Under-the-hood stability improvements Still 100% offline. No accounts, no ads, no tracking.

New in 1.1.0 — Neural Voice Conversion Go beyond effects and truly become someone else. Our new AI voices reshape your actual timbre, not just the pitch: • Narrator, Deep Giant, Young Man, Soft Lady, Mature Woman • Works on both recordings and typed text • Runs 100% on-device — your voice never leaves your phone Also in this release: • Redesigned voice picker with character portraits • Smoother, more natural text-to-speech • Stability and performance improvements Still 100% offline. No accounts, no ads, no tracking.

Fix BUG