Local ASR model

Phonon-2

Q: Can Phonon-2 run locally?

Phonon-2 is listed by LocalClaw as a local ASR option. Hardware fit depends on runtime, model size and backend support.

Catalogue summary: English on-device ASR model derived from Parakeet TDT 0.6B v3 with quantization-aware packed weights. The official Hugging Face release ships a 164 MB package, word-level JSON timestamps, MLX and Core ML paths for Apple Silicon, plus CPU/GPU Docker engines and the `fermion-research` local CLI.

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

CPU friendlyspeech-to-text transcription1 languagesCC-BY-4.0 weights; Apache 2.0 code

Choose an app

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

Compare speech models

Open model filesRecommended verified starting point Open in UnslothLaunches Unsloth Desktop on this model Open Python packageVerified package page without exposing a command Open on Hugging FaceFiles, licence and model card Check 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

9.1/10

Catalogue speed

9.7/10

Model size

164 MB packed weights; sub-1GB ASR download

Voices

N/A (ASR: outputs English transcripts and optional word timestamps)

Can Phonon-2 run locally?

Phonon-2 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.

CC-BY-4.0 weights; Apache 2.0 code license. Still verify upstream usage notes before shipping.

Upstream source

realtimelow-latency

Audio profile

Cat. quality

9.1

Cat. speed

9.7

Audio

9.3

Best fit

Phonon-2 is best for offline transcription, speech indexing and local voice pipelines.

Hardware: cpugpuapple

Model details

Type

Local ASR model

Family

phonon

Latency

ultra-low

Formats

mlxcoremlpythondocker

Languages

Context

English-only speech-to-text, 164 MB download, Open ASR Leaderboard-style WER card, `phonon transcribe` and `fermion transcribe phonon-2` commands, Core ML companion export and CPU/CUDA Docker engines; upstream benchmarks include M5 MacBook Air, Zen 5 CPU and H100 batch runs but no fixed RAM floor

Install locally

Check runtimeConfirm the backend supports mlx, coreml, python, docker on your machine.

Open recommended setupUse the app and model links above. LocalClaw does not expose a terminal command.

Test locallyRun a short private audio prompt before moving into production workflows.

Good for

speech-to-text transcription
CPU friendly local workflows
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

CompareBrowse all TTS models Local AIBrowse LLM models macOS appGet LocalClaw