remiqora
Free, open-source local AI music studio: ACE-Step 1.5 and YuE2-3B text-to-music, stem separation, an AI arranger that adds parts to your track, a multitrack DAW
What is remiqora?
Free, open-source local AI music studio: ACE-Step 1.5 and YuE2-3B text-to-music, stem separation, an AI arranger that adds parts to your track, a multitrack DAW, MIDI transcription and LoRA training.
Overview of remiqora
What it is
Remiqora is a local, GPU-powered music generation and production studio. It provides a single interface for ACE-Step 1.5, YuE2-3B, and Stable Audio 3, along with a built-in multitrack DAW, an AI arranger that writes new parts into a track, and a feature called “From an idea” where a local language model writes the prompt from one phrase. It is free, open-source, and supports stem separation, MIDI transcription, and LoRA training.
Key features
- ACE-Step 1.5: Fast generation from text/style tags, covers, section repainting, and extracting/adding parts on top of a reference track.
- YuE2-3B: Full-length track generation with CoT score planning (a symbolic ABC plan before the audio).
- Stable Audio 3: Instrumentals and sound effects from a description in seconds.
- From an idea: One phrase in, a local language model (Qwen3.5) writes the prompt for ACE-Step, YuE2, or Stable Audio 3.
- LoRA training: Dataset → auto-labeling → preprocessing → training → export for ACE-Step fine-tuning, all in the browser.
- Demucs: Splits any track into 4 stems: vocals, drums, bass, other.
- Built-in DAW: Multitrack timeline editor with effects rack, auto-BPM, time-stretch, WAV/MP3 export, and Reaper project / DAWproject export.
- AI arranger: Adds a part (drums, bass, strings, keys, vocals, etc.) on top of chosen lanes, repaints a section, continues the track, or covers it in another style.
Use cases
- Generating music from text descriptions or style tags.
- Separating a mixed track into individual stems (vocals, drums, bass, other).
- Transcribing audio (full mix or a single stem) into MIDI notes.
- Adding AI-generated instrumental or vocal parts to an existing track.
- Fine-tuning ACE-Step on a custom voice or style using LoRA training.
- Assembling and mixing multiple tracks or stems into a final composition using the built-in DAW.
Getting started
See the repository README for installation instructions.
License and pricing
The project is licensed under the MIT License and is free and open-source.
Based on the project's public README and repository metadata.