spb/tools-spboucher-ai

1{2  "slug": "mlx-whisper",3  "name": "MLX Whisper",4  "aliases": ["mlx-whisper", "mlx_whisper"],5  "vendor": { "name": "Apple (MLX contributors)", "slug": "apple", "website": "https://github.com/ml-explore" },6  "homepage_url": "https://github.com/ml-explore/mlx-examples/tree/main/whisper",7  "docs_url": "https://pypi.org/project/mlx-whisper/",8  "repo_url": "https://github.com/ml-explore/mlx-examples",9  "license_url": "https://github.com/ml-explore/mlx-examples/blob/main/LICENSE",10  "product_type": "library",11  "product_status": "active",12  "summary": "Port of OpenAI's Whisper speech-recognition models to MLX, offering a CLI and Python API for local transcription on Apple Silicon",13  "description": "MLX Whisper is the Whisper example from Apple's `mlx-examples` repository packaged as a PyPI library (`mlx-whisper`, MIT license). It runs OpenAI's open Whisper speech-recognition models on Apple Silicon through MLX rather than PyTorch, using pre-converted weights hosted in the `mlx-community` organization on the Hugging Face Hub or models you convert yourself.\n\nThe `mlx_whisper` command transcribes an audio file, or audio piped on standard input, into text, SRT, VTT or JSON output, with options for model choice, language and translation. The Python `transcribe` function returns segments and can produce word-level timestamps. ffmpeg must be installed for audio decoding. It is a convenient choice for researchers and developers who want fast local transcription of lectures, interviews or recordings without sending audio to a cloud service.\n\nLimits are those of Whisper itself (accuracy varies by language and audio quality, no built-in speaker diarization) plus the Apple Silicon requirement; it is a focused example package rather than a full transcription product.",14  "capabilities": ["Local transcription of audio files or stdin streams", "Multiple output formats including text, SRT, VTT and JSON", "Word-level timestamps", "Choice of Whisper model sizes from the Hugging Face Hub", "Python API alongside the command-line tool"],15  "workflows": ["Transcribe recorded lectures or meetings on a Mac", "Batch-generate subtitles for a video archive", "Embed speech-to-text in a local Python pipeline"],16  "advantages": ["No cloud dependency; audio stays on the machine", "Simple pip install and single command", "MIT license"],17  "limitations": ["Apple Silicon only", "No speaker diarization", "Requires ffmpeg and model downloads", "Maintained as an example package rather than a standalone product"],18  "prerequisites": ["Mac with Apple Silicon", "Python 3.8 or newer", "ffmpeg"],19  "audiences": ["Researchers transcribing interviews and lectures", "Developers adding local speech-to-text to Mac tools"],20  "alternatives": ["whisper", "mlx"],21  "categories": [{ "slug": "apple-silicon-ml", "primary": true }, { "slug": "speech-vision-multimodal" }],22  "tags": ["apple-silicon", "macos", "speech-recognition", "transcription", "open-source", "python", "cli"],23  "use_cases": ["run-local-models", "transcribe-audio"],24  "operating_systems": ["macos"],25  "execution_modes": ["local"],26  "interfaces": ["cli", "library"],27  "deployment_models": ["desktop"],28  "integrations": ["MLX", "Hugging Face Hub", "ffmpeg"],29  "data_formats": ["MP3", "WAV", "M4A (via ffmpeg)", "TXT", "SRT", "VTT", "JSON"],30  "languages": ["Python"],31  "source_model": "open-source",32  "license_spdx": "MIT",33  "pricing_model": ["open-source", "free"],34  "self_hostable": true,35  "sources": [36    { "url": "https://github.com/ml-explore/mlx-examples/tree/main/whisper", "type": "official_site", "title": "mlx-examples/whisper README", "supports": ["name", "summary", "capabilities", "prerequisites"], "checked_at": "2026-11-02" },37    { "url": "https://pypi.org/project/mlx-whisper/", "type": "package_registry", "title": "mlx-whisper on PyPI", "supports": ["product_status", "prerequisites", "license_spdx"], "checked_at": "2026-11-02" },38    { "url": "https://github.com/ml-explore/mlx-examples/blob/main/LICENSE", "type": "license_page", "supports": ["license_spdx", "source_model"], "checked_at": "2026-11-02" }39  ],40  "confidence": "high",41  "discovery_method": "editorial-research",42  "depth": "standard",43  "editorial_status": "reviewed",44  "visibility": "public",45  "featured": false,46  "author": "batch-C"47}48