Registry / ai-ml / mlx-embeddings

mlx-embeddings

library0.1.0pypypiunverified

MLX-Embeddings is a Python package for generating text and vision embeddings locally on Apple Silicon Macs using MLX. It supports models like BERT, CLIP, and SigLIP, optimized for Apple's Metal Performance Shaders. Version 0.1.0 is the initial release, with modular backends.

ai-ml

pip install mlx-embeddings

github ↗homepage ↗

INSTALL

IMPORT

SIG · MLX-EMBEDDINGS

mlx-embeddings

ai-mlpythonv0.1.0

harness data pending

Install & Compatibility

Where this runs

No compatibility data collected yet for this library.

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

TextEmbedding

✓ from mlx_embeddings import TextEmbedding

✗ from mlx_embeddings import TextEmbedding

Quickstart example: create a text embedding model and encode a sentence.

from mlx_embeddings import TextEmbedding import os # Create a text embedding object for a BERT model embedder = TextEmbedding(model_name="bert-base-uncased") # Generate embeddings for a sentence embeddings = embedder.embed(["Hello world"]) print(embeddings.shape) # For vision models: # from mlx_embeddings import ImageEmbedding # image_embedder = ImageEmbedding(model_name="openai/clip-vit-base-patch32") # image_embeddings = image_embedder.embed(["path/to/image.jpg"])

Debug

Known issues

gotchaMLX-Embeddings only works on Apple Silicon (M1/M2/M3) Macs. Execution on Intel Macs or other platforms will fail with a runtime error.

fix

Ensure you are on a Mac with Apple Silicon (ARM64).

affects: 0.1.0

gotchaThe package auto-downloads model weights from Hugging Face. The first run for each model will download several hundred MB to GB of data, which may fail without an internet connection or sufficient disk space.

fix

Pre-download models using huggingface_hub or ensure stable internet on first use.

affects: 0.1.0

gotchaVision models (ImageEmbedding) require Pillow. Without it, importing or using ImageEmbedding may raise ModuleNotFoundError.

fix

Install optional dependency: pip install mlx-embeddings[vision] or pip install pillow.

affects: 0.1.0

Upgrade

Version history

0.1.0latest on PyPI · released Mar 24, 2026

Audit

Dependencies

mlxrequiredRequired for running models on Apple Silicon.

pillowoptionalRequired for image processing in vision models.

Agent activity

15 hits · last 30 days

node

Amazon

Resources

githubhttps://github.com/Blaizzy/mlx-embeddings ↗

homepagehttps://github.com/Blaizzy/mlx-embeddings ↗