LocalClaw/TTS/ASR4All Small
Local ASR model

ASR4All Small

Catalogue summary: FUTO's efficiency-focused English ASR checkpoint for private edge transcription. The 28.1M-parameter small tier ships downloadable safetensors plus ONNX and ExecuTorch XNNPACK exports, supports low-latency streaming with punctuation, capitalization and error correction, and is benchmarked for CPU, mobile, Raspberry Pi, OpenVINO and CUDA paths.

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

Edge readyspeech-to-text transcription1 languagesFUTO Model Weights License (Restricted Use) 1.0
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 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
8.9/10
Catalogue speed
9.7/10
Model size
28.1M params; 1.31GB repo with safetensors, ONNX and ExecuTorch exports
Voices
N/A (streaming ASR: outputs punctuated text)

Can ASR4All Small run locally?

ASR4All Small 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.

FUTO Model Weights License (Restricted Use) 1.0 license. Review upstream restrictions before commercial use.

Upstream source
streamingrealtimelow-latency

Audio profile

Cat. quality
8.9
Cat. speed
9.7
Audio
9.2

Best fit

ASR4All Small is best for offline transcription, speech indexing and local voice pipelines.

Hardware: cpugpuedge

Model details

Type
Local ASR model
Family
asr4all
Latency
ultra-low
Formats
transformerssafetensorsonnxexecutorch
Languages
en
Context
28.1M params, English-only, 16 kHz audio, configurable 320 ms to long-form streaming tiers, bundled PCEC and ultralight VAD; official benchmarks include Pixel 4, Pixel 10, Raspberry Pi 4, x86 OpenVINO and CUDA batch throughput, but no fixed RAM floor is published

Install locally

01
Check runtimeConfirm the backend supports transformers, safetensors, onnx, executorch 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
  • Edge ready local workflows
  • streaming, realtime, 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

NVIDIA Parakeet TDT 0.6B v2 Local ASR model · Q 9.4 · Speed 10 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 X-Square Robot X2Streaming-ASR 4B 1009 Local ASR model · Q 9.2 · Speed 9.5 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
CompareBrowse all TTS models Local AIBrowse LLM models macOS appGet LocalClaw