mlx-whisper
OpenAI Whisper on Apple silicon with MLX and the Hugging Face Hub
Decision gist · record as of 2026-08-14
Yes, if you are on Apple silicon and need local speech recognition. The package is actively maintained, has low install friction, and MIT licensing poses no restrictions. Verify that torch and mlx dependencies resolve cleanly on your target hardware, and ensure ffmpeg is available before installation.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires ffmpeg to be installed on the system (e.g., `brew install ffmpeg` on macOS) and torch/mlx dependencies which are platform-specific.
- Low install friction with a pure Python wheel.
- Active maintenance with recent commits and 8879 repository stars.
License · maintenance · safety
MIT (permissive) — MIT license permits commercial and private use with minimal restrictions, making it suitable for most projects.
last release 2025-08-29 (350 days) · last repo commit 2026-04-06 · 8,879 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 336,970 downloads/mo, #7,453 on PyPI
Alternatives
Verify before relying
pip install mlx-whisper
import mlx_whisper
text = mlx_whisper.transcribe("audio_file.mp3")["text"]- Whether MLX and torch dependencies install smoothly on non-Apple-silicon systems or if they are strictly required for Apple hardware.
- Performance characteristics and latency for different model sizes on typical Apple silicon devices.
- Whether pre-converted models from Hugging Face Hub cover all use cases or if manual conversion is often needed.
What it is and what it does
mlx-whisper brings OpenAI's Whisper speech recognition to Apple silicon Macs by wrapping the models in MLX, a machine learning framework optimized for Apple hardware. It offers both a command-line interface for simple transcription and a Python API for programmatic use. The package can load models from the Hugging Face Hub or local paths, supporting models at various parameter scales and offering features like word-level timestamps.
The package depends on torch, mlx, numpy, scipy, and several utility libraries for audio processing and model management. It requires ffmpeg as a system dependency for audio handling. Transcription is performed locally on-device, which means no network calls to external APIs and full control over model selection and quantization.
Use it for
- Transcribe audio files to text on macOS without sending data to external APIs.
- Build command-line tools that convert speech recordings (MP3, WAV, etc.) to text files.
- Integrate speech-to-text into Python applications with control over model size and precision.
- Process audio streams piped from other programs directly into transcription.
- Generate word-level timestamps for audio segments to enable precise subtitle generation.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, if you are on Apple silicon and need local speech recognition.
The package is actively maintained, has low install friction, and MIT licensing poses no restrictions. Verify that torch and mlx dependencies resolve cleanly on your target hardware, and ensure ffmpeg is available before installation.
Install
mlx-whisper on PyPI
Before you install
Low install friction with a pure Python wheel. Active maintenance with recent commits and 8879 repository stars. Requires ffmpeg as a system dependency and nine runtime packages including torch, mlx, and numpy.
Requires ffmpeg to be installed on the system (e.g., `brew install ffmpeg` on macOS) and torch/mlx dependencies which are platform-specific.
License in practice
MIT license permits commercial and private use with minimal restrictions, making it suitable for most projects.
Quickstart
pip install mlx-whisper
import mlx_whisper
text = mlx_whisper.transcribe("audio_file.mp3")["text"]Verify before relying
- Whether MLX and torch dependencies install smoothly on non-Apple-silicon systems or if they are strictly required for Apple hardware.
- Performance characteristics and latency for different model sizes on typical Apple silicon devices.
- Whether pre-converted models from Hugging Face Hub cover all use cases or if manual conversion is often needed.
Package facts
| License | MIT permissive |
| Python support | Supports the current Python release >=3.8 |
| Install friction | Low. Pure-Python wheel |
| Runtime dependencies | 9 packagesmlxnumbanumpytorchtqdmmore-itertoolstiktokenhuggingface_hubscipy |
| Maintenance | Actively maintained 350 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 336,970 / month, #7,453 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
Evidence: mlx_whisper-0.4.3-py3-none-any.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “speech recognition apple silicon”
- mlx-whisperRuns OpenAI's Whisper speech recognition models on Apple silicon…
- mlx-audioMLX-Audio provides text-to-speech, speech-to-text, and…
- pyobjc-framework-SpeechProvides Python bindings to macOS's native Speech framework, enabling…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also mlx-lm · openai-whisper · mlx-audio · mlx · mlx-vlm · whisperx · faster-whisper · SpeechRecognition · whisper-normalizer · whisper-timestamped