UtaForge AI is a free, open-source local tool for creating VOICEVOX singing covers and lyric videos from audio and lyrics. It analyzes a source vocal’s pitch, volume, and lyric timing, then transfers those elements to a VOICEVOX voice. Lyrics can be supplied as text, Markdown, or LRC, with line and mora timing edited on a waveform. The project creates subtitled MP4 videos in vertical 9:16 or widescreen 16:9 formats. Visuals may use PSD-based lip-sync, blinking, facial expressions, or a still image. Exports include audio, video, VOICEVOX .vvproj projects, and UtaFormatix .ufdata files. A browser-based Studio and command-line interface are available. Processing stays on the local computer and does not require external web services. VOICEVOX ENGINE and ffmpeg are required; audio separation can use audio-separator or Demucs. It supports macOS 13 or later and Windows 11 with Python 3.11 or 3.12. The project is alpha and uses the MIT License.
Who it is for
UtaForge AI suits creators making VOICEVOX covers or lyric videos on macOS or Windows. Users should have rights to the music, lyrics, voices, images, characters, and generated works they use.
What is good
- Free and open source
- Runs processing locally
- Supports text, Markdown, and LRC lyrics
- Exports audio, video, and project files
- Provides browser Studio and command-line interfaces
What to know first
- Project status is alpha
- Requires VOICEVOX ENGINE and ffmpeg
- Supports macOS 13+ or Windows 11
- Output quality and formats may change
Verdict
UtaForge AI provides local tools for transferring vocal performance details to VOICEVOX voices and producing lyric videos. Its alpha status and required companion software are important considerations for setup and ongoing use.
Compared on AI song cover generators
- Vocal upload
- Yes
- Instrumental upload
- Yes
- Export formats
- MP3
