LocalClaw/TTS/X2Streaming-ASR 4B 1009
Local ASR model

X2Streaming-ASR 4B 1009

Catalogue summary: Apache-licensed append-only streaming ASR model for Chinese, English and Chinese-English code-switching. It starts from Voxtral Mini Realtime, replaces fixed-delay decoding with a learned wait-or-emit policy, ships an 8.86 GB safetensors checkpoint and runs locally through the official Transformers or vLLM inference stack.

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

GPU recommendedspeech-to-text transcription3 languagesApache 2.0
Choose an app

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

Compare speech models
Open model filesRecommended verified starting pointOpen 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
9.2/10
Catalogue speed
9.5/10
Model size
8.86 GB safetensors; inherits the Voxtral 4B single-GPU 16GB+ class runtime
Voices
N/A (streaming ASR: outputs append-only text)

Can X2Streaming-ASR 4B 1009 run locally?

X2Streaming-ASR 4B 1009 can run locally for offline speech-to-text. Use the verified setup options on this page; no terminal command is required to choose the right path.

Apache 2.0 license. Still verify upstream usage notes before shipping.

Upstream source
streamingrealtimemultilinguallow-latency

Audio profile

Cat. quality
9.2
Cat. speed
9.5
Audio
9.3

Best fit

X2Streaming-ASR 4B 1009 is best for offline transcription, speech indexing and local voice pipelines.

Hardware: gpu

Model details

Type
Local ASR model
Family
x2streaming
Latency
ultra-low
Formats
transformersvllmsafetensors
Languages
zh, en, cs
Context
4B streaming ASR, 16 kHz mono input, about 80 ms audio frames, adaptive and offline modes, zh/en/cs language flag required for each run

Install locally

01
Check runtimeConfirm the backend supports transformers, vllm, safetensors 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

  • speech-to-text transcription
  • GPU recommended local workflows
  • streaming, realtime, multilingual

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

NVIDIA Parakeet TDT 0.6B v3 Local ASR model · Q 9.5 · Speed 9.7 NVIDIA Nemotron 3.5 ASR Streaming 0.6B Local ASR model · Q 9.4 · Speed 9.3 Alibaba Cloud (Qwen Team) Qwen3-ASR Local ASR model · Q 9.5 · Speed 9 Cactus Compute Whistle Local ASR model · Q 9 · Speed 9.8 Mistral AI Voxtral Mini 4B Realtime 2602 Local ASR model · Q 9.2 · Speed 9.4 OpenAI Whisper v3 Turbo Local ASR model · Q 9.1 · Speed 9.5 Audar AI Labs Audar-ASR-V1-Flash Local ASR model · Q 9.1 · Speed 9.4 FunAudioLLM Fun-ASR-MLT-Nano-2512 Local ASR model · Q 9.2 · Speed 9.2
CompareBrowse all TTS models Local AIBrowse LLM models macOS appGet LocalClaw