Open your video
Bring in your recording. Speech and pauses are analyzed locally on your computer.
Clean up pauses, retakes and filler words. Add subtitles, translate them and dub in your own voice. Right on your Mac.
Focus on what you have to say. NU AI finds the extra parts in your recording. You make the final call.
Bring in your recording. Speech and pauses are analyzed locally on your computer.
See every proposed cut on the transcript and timeline. Keep it, remove it, or make your own.
Get a smoother MP4. Add subtitles and extra language tracks when your story needs them.
Cleanup, subtitles and export in one place. These are real screenshots from the NU AI app.
Pauses, false starts, filler sounds, accidental repeats and frozen footage where nobody is speaking. See why each cut was suggested and decide what stays.
Transcription, speech detection and voice cloning run on your computer. Optional AI review and translation send only transcript text to the provider you choose.
Bring your own API provider or use a local model with Ollama or LM Studio.
Download for free and try your own recording. Basic cleanup needs no AI provider or API key.
Open the DMG and drag the nuai app into your Applications folder.
If you use Homebrew, run the command below.
brew install ffmpegChoose a model in Settings → Transcription inside the app. The recommended model is about 547 MB.
The app is not Apple-notarized yet. See the installation guide for the first launch. Intel Mac users can build from source.
Yes. NU AI is free and open source under the MIT license. If you use your own API provider, that provider’s usage charges may apply separately.
The ready-to-install download is currently for Apple Silicon Macs. Windows and Linux packages are planned. Intel Mac users can build the app from source.
No. Local transcription, pause detection and basic cleanup work without an AI provider. AI review, target-length shortening and translation use your chosen provider.
Video and audio files are processed locally. If you enable optional cloud AI features, only transcript text is sent to your chosen provider. You can also use a local model.
Beta dubbing takes a voice reference from the speaker in your video and uses it to speak translated captions while preserving the background. The optional local voice engine requires about a 4.5 GB download. Only clone voices you own or have permission to use.
Yes. In Review mode you can keep or remove each suggested cut, select words to make your own cuts and preview the result before exporting.