Description
Record any voice for 5 seconds. Clone it. Type anything and hear it spoken back — in 10+ languages, with built-in on-device translation. Everything happens on your device. No cloud, no uploads, completely private.
VoiceNPC is your pocket voice copycat. Powered by the Qwen3-TTS neural network running locally via Apple's MLX framework, it captures a voice's unique fingerprint and recreates it with stunning accuracy. Your voice data never leaves your device.
Works on iPhone, iPad, and Mac with Apple Silicon.
What you can do:
- Clone any voice from a 5–15 second recording or import from a video
- Generate speech in English, Chinese, Japanese, Korean, French, German, Spanish, Russian, Italian, Portuguese, and more
- Translate text on-device and have your cloned voice speak it back — type in one language, hear it in another
- Pro: pick multiple target languages and generate them all in a single pass
- Preview and edit translations before audio synthesis
- Create multi-voice projects — assign different voices and languages to each section for dialogues, audiobooks, and podcasts
- Fine-tune generation quality with style presets (Stable, Natural, Expressive) or dial in custom settings
- Export as WAV or share voice fingerprints (.voicenpc files) — let others generate speech with your voice via AirDrop, Messages, or any app
- Create unlimited voice clones on every plan
Built for creators:
- YouTubers — dub your videos in multiple languages with your own voice, no separate translation step
- Podcasters — produce multi-speaker episodes without extra recording sessions
- Audiobook narrators — assign a unique voice to each character
- Voiceover artists — share your voice fingerprint with clients so they can generate lines on demand
- Game developers — generate localized NPC dialogue at scale
Why it's different:
- 100% on-device — works offline, no internet needed after install (translation included)
- Zero data uploaded — your recordings and source text stay in the app sandbox
- No cloud processing — your voice never touches a server
- Powered by Apple Silicon — fast, efficient, private
How energy works:
- 1 energy ≈ 100 characters of text (about 1–2 sentences)
- 25 energy included daily — enough for ~5 full generations
- Watch a short video for 25 bonus energy (up to 8 per day)
- Pro subscribers get unlimited generations with no ads
Go Pro for unlimited:
- Monthly US$2.99 (less than US$0.10/day)
- Annual US$29.99 (2 months free)
- Lifetime US$59.99 (pay once, yours forever)
Requires iPhone 13 or later, iPad mini (6th gen) or later, iPad Air (5th gen) or later, iPad (11th gen) or later, or any Apple Silicon Mac.
Terms of Use: Standard Apple Terms of Use (EULA) [https://www.apple.com/legal/internet-services/itunes/dev/stdeula/]
What's new (v1.3.1)
What's new in 1.3.1:
• In-app translation — type in one language and have your cloned voice speak it in another, all on-device
• Pro: pick multiple target languages and generate them all in a single pass
• Preview and edit translations before audio synthesis
• Sharper voice-recording transcription powered by Apple's latest on-device speech engine
• More reliable translation pack setup with clearer error feedback
• Bug fixes and stability improvements