mlx-smolvla is an open-source Python package providing native MLX inference, serving, and training tools for SmolVLA models on Apple Silicon. It is intended for developers and robotics researchers building local vision-language-action workflows.
mlx-smolvla sits in PulseGate's Inference & model serving category. It focuses on running and serving SmolVLA vision-language-action models efficiently on Apple Silicon. mlx-smolvla is an open-source project aimed at developers and robotics researchers using Apple Silicon. mlx-smolvla is open source under the Apache-2.0 license. mlx-smolvla is available on the command line and API, and it can be self-hosted.
Behind mlx-smolvla is Daniiar Abdiev, and it first shipped in 2026. Development happens publicly on GitHub with 139 commits in the last 90 days. Key capabilities include MLX inference, model serving, and model training.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do