AI tools, scored for your job
Get paid to train AI
🤝 You're in the Free & Open Source library — community-built tools you can use without paying. A different concept from our scored AI directory.
Home › Free & Open Source › Audio & Music › MOSS-Audio-Tokenizer
✦ Free & open sourceAudio & Music

MOSS-Audio-Tokenizer

A 1.6B causal Transformer audio tokenizer for speech, sound, and music.

MOSS-Audio-Tokenizer

Unified discrete audio tokenization with semantic alignment and streaming support.

Free forever · open source · Apache-2.0
✓ Needs IT✓ Browser-based✓ Real-time collaboration
6
ZEKAI
Powerful 1.6B audio tokenizer
Free & open source · Audio & Music
Ease of setup4
Usefulness8
Maintenance7
Community5
★ 268
GITHUB
Jun 2026
UPDATED
Apache-2.0
LICENSE
🛠 Don't want to self-host? We can set it up →
⚡ Quick answer

MOSS-Audio-Tokenizer is a 1.6-billion-parameter causal Transformer audio tokenizer designed for next-generation audio foundation models. It suits researchers and audio engineers building advanced text-to-speech, speech recognition, or generative audio systems.

Runs on
Cross-platform
Hosting
Self-host
Languages
Python
License
Apache-2.0

What it is

It compresses raw audio into a low frame rate of 12.5Hz using a pure Transformer architecture built entirely from causal blocks without CNNs. Trained on 3 million hours of diverse data, it jointly optimizes encoder, decoder, RVQ, discriminator, and an LLM for semantic alignment in a single pipeline.

Who it's for

It suits machine learning researchers and audio developers building custom speech or music generation pipelines. It does not suit casual users looking for a ready-to-use desktop audio editor or simple plugin.

Key features

Pure Transformer Architecture

Built entirely from causal Transformer blocks without CNNs to support low-latency streaming.

Variable Bitrate RVQ

Employs a 32-layer Residual Vector Quantizer supporting bitrates from 0.125kbps to 4kbps.

Semantic-Acoustic Alignment

Produces discrete tokens that achieve both lossless-quality reconstruction and high-level semantic alignment.

Multiple Backends

Provides PyTorch, ONNX Runtime, and TensorRT inference backends for flexible deployment.

How to start

Up and running in 3 steps

1
Clone the official GitHub repository and install the required dependencies
2
Download model weights from Hugging Face or ModelScope
3
Load the model using PyTorch or export/run via the provided ONNX Runtime scripts
Real talk

Our honest verdict

What's great

  • Trains end-to-end from scratch without relying on pretrained teachers like HuBERT or Whisper
  • Handles speech, sound effects, and music across a unified 1.6B parameter model
  • Offers alternative ONNX and TensorRT runtimes for deployment outside PyTorch
  • Supports extreme compression down to 12.5Hz frame rate with variable bitrates

Where it falls short

  • Demands significant compute resources to run and deploy a 1.6B parameter model
  • Requires technical machine learning expertise to integrate into custom downstream pipelines
  • Not a turnkey end-user application for immediate audio production tasks
Cost

Pricing & license

Free under its open-source license, with the only cost being the hardware infrastructure required to host and run a 1.6B parameter model.

Paid work

Open remote AI-training roles

897 open remote AI-training roles, updated November 17, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.

See all open roles →
Compare

Alternatives

🎁 Free & open-source alternatives

These reviewed AI tools do the same job instantly — from our scored directory.

Get more

Use it, or have us set it up

📄 Free PDF: MOSS-Audio-Tokenizer for Audio & Music
A plain-language guide to getting your first result. Sent to your inbox.
🔔 64 people want a full video tutorial — add your vote
🛠 Don't want to set it up yourself?
We'll install, host & configure MOSS-Audio-Tokenizer for you — and train your team.
Get it done for you ↗
Questions

FAQ

What data was MOSS-Audio-Tokenizer trained on?

It was trained on 3 million hours of diverse audio data covering speech, sound effects, and music.

Can I run this model without PyTorch?

Yes, the repository provides ONNX Runtime and TensorRT inference backends for deployment without PyTorch.

Radar

Latest AI news

✉️ The weekly free toolkit
The best new free & open-source tools for Audio & Music — one short email a week, in your language.
Free · 1 email/week · unsubscribe anytime
Get paid to train AI →

See Zekai first in Google