OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Magenta RealTime 2

Magenta RealTime 2 is an open-weights music-generation model with Python and C++ inference software and example applications. It supports prompt-based inference and is designed for real-time music streaming on specified Apple Silicon hardware.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Magenta RealTime 2 — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when you want a local model and developer components for music generation, especially for real-time workflows on Apple Silicon.

Who it is for

  • Musicians exploring generated material for live or studio workflows.
  • Python developers running music-generation inference.
  • C++ developers building music applications or integrations such as a DAW plugin.

Documented uses

  • Generate a short music clip from a text prompt.
  • Run music-generation inference locally through the Python library.
  • Build an application around the C++ inference engine or use the provided DAW plugin example.
Before you choose

Requirements & limits

  • The real-time path requires Apple Silicon; the repository lists offline Python inference on Apple Silicon or NVIDIA GPUs.
  • The 230-million-parameter small model is listed for real-time use on any Apple Silicon Mac, while the 2.4-billion-parameter base model requires a Pro Max chip for real-time streaming.
  • The project offers separate Python and C++ inference components, plus example apps and plugins.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Generate a short prompt-based clip with the MLX quickstart

Set up the Python inference package and generate four seconds of music on Apple Silicon.

Before you start

  • An Apple Silicon Mac for the documented MLX quickstart.
  • Python 3.12 and uv.
  • Internet access to install the package and download model resources.

Steps

  1. Create and activate a Python 3.12 virtual environment with uv.

  2. Install the MLX package extra with the command below.

    Documented command
    uv pip install "magenta-rt[mlx]"

  3. Run mrt models init to download the style and codec resources.

  4. Run mrt models download to download the streaming model.

  5. Generate a four-second clip with mrt mlx generate --prompt "disco funk" --duration 4.0 --model=mrt2_base.

Expected outcome

The command requests a four-second music generation using the prompt “disco funk” and the base model.

Guide limitations

  • Real-time streaming with the base model is listed for Pro Max chips; the smaller model is the documented alternative for broader Apple Silicon real-time support.
  • The guide describes inference setup and a generation command, not a guarantee of runtime performance on a particular device.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Model inference and serving

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
Magenta RealTime 2Generate a short music clip from a text prompt.
Pulse Of MotionEstimate a video's physical frame rate from its visual dynamics.LinuxThe repository is public. Its README describes a released checkpoint that downloads from Hugging Face on first inference, with a manual download option. The strongest internal model is not fully released; the README directs interested users to contact the project about access. No license information is provided in the repository evidence.
Neural Image AssessmentTrain the model using AVA images and rating annotations.Cloud, Linux, macOS, Windows
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
11 October 2026
Original website title
GitHub - magenta/magenta-realtime: Magenta RealTime 2: An Open-Weights Live Music Model · GitHub · Official homepage title · retained in its source language
How we review sourcesSuggest a correction