AceForge is a local-first AI music workstation for macOS Silicon powered by ACE-Step

Features:

• .100% Local (only needs to download models once)
• Music Generation with ACE-Step prompts
  • Use Stem separation to rebalance vocals vs instrumentals
  • Use existing Audio as reference (optional)
  • Train ACE-Step LoRAs from your own datasets
    • Mass-create _prompt.txt / _lyrics.txt files
    • Auto-tag datasets using MuFun-ACEStep (experimental)
• Stem Splitting using Demucs for high-quality audio separation
• Voice Cloning TTS using XTTS v2
• MIDI Generation using basic-pitch for audio-to-MIDI transcription
• Embedded Music Player to explore generation catalog
• Manage and reuse prompt presets

github.com/audiohacking/AceForge