LocalClaw/TTS/Mizar-159M
Local audio-language model

Mizar-159M

Catalogue summary: Compact audio-language model for local audio question answering. The 159M-parameter release combines a frozen CED-Small audio encoder, a frequency-merging mapper and SmolLM2-135M, ships complete checkpoint assets and answers English questions about the first 20 seconds of an audio file through the official PyTorch plus vLLM pipeline.

Repository editorial metadata; verify comparative claims in the linked upstream material.

CPU friendlyaudio-language question answering1 languagesBSD 3-Clause Clear
Choose an app

Start with the recommended app, or choose another supported option.

Compare speech models
Open model filesModel weights and official local setupOpen in UnslothLaunches Unsloth Desktop on this modelOpen official projectApplication source and guided setupOpen on Hugging FaceFiles, licence and model cardCheck my machineOptional: verify the fit with your saved hardware

Desktop app links require the app to be installed. If nothing opens, LocalClaw will show app-download and model-file fallbacks.

Catalogue quality
8.7/10
Catalogue speed
8.8/10
Model size
159.3M params; checkpoint plus bundled CED-Small and SmolLM2 assets
Voices
N/A (audio-language question answering)

Can Mizar-159M run locally?

Mizar-159M can run locally for audio-language question answering over supplied recordings. Use the verified setup options on this page; no terminal command is required to choose the right path.

BSD 3-Clause Clear license. Still verify upstream usage notes before shipping.

Upstream source
audio-languagelow-latency

Audio profile

Cat. quality
8.7
Cat. speed
8.8
Audio
8.7

Best fit

Mizar-159M is best for local audio question answering and lightweight audio understanding experiments.

Hardware: cpugpu

Model details

Type
Local audio-language model
Family
mizar
Latency
low
Formats
pytorchsafetensorsvllm
Languages
en
Context
Official code prepares audio prefixes on CPU or CUDA, then uses vLLM GPU generation; BSD 3-Clause Clear weights include the final checkpoint, CED-Small and SmolLM2 assets

Install locally

01
Check runtimeConfirm the backend supports pytorch, safetensors, vllm on your machine.
02
Open recommended setupUse the app and model links above. LocalClaw does not expose a terminal command.
03
Test locallyRun a short private audio prompt before moving into production workflows.

Good for

  • audio-language question answering
  • CPU friendly local workflows
  • audio-language, low-latency

Watch before shipping

  • Validate pronunciation, latency and artifacts with your own voice samples.
  • Review the upstream license and acceptable-use notes.
  • Benchmark on your target CPU, Apple Silicon or GPU setup.

Related TTS and speech models

jamiepine / Community Voicebox Local speech app · Q 9 · Speed 9.5 OpenLLM France / LINAGORA Luciole 8B Audio 1.0 Local audio-language model · Q 9.1 · Speed 6.7 hexgrad Kokoro TTS Local TTS model · Q 9.2 · Speed 9.8 Samuel Vitorino / HALO Research Sopro V2 Turbo Local TTS model · Q 9.2 · Speed 9.4 Canberk Aslan EMA Lightning Local TTS model · Q 9 · Speed 9.8 Audio8 Audio8 TTS Preview 0.6B Local TTS model · Q 9.3 · Speed 8.9 dots-studio Dots TTS MF 2-step Local TTS model · Q 9.4 · Speed 8.7 Kyutai Moshi Local TTS model · Q 9 · Speed 9.5
CompareBrowse all TTS models Local AIBrowse LLM models macOS appGet LocalClaw