ml-exploreml-exploreOfficialยท1 Agent Skills Included

mlx-swift-lm

Run local LLMs and vision models on Apple Silicon with Swift

Runs large language models and vision-language models locally on Apple Silicon using Swift and MLX. Handles streaming chat, tool calling, embeddings, LoRA fine-tuning, and memory-safe concurrent inference without cloud APIs. Gives developers ready-made Swift APIs and reference guides to build fast, private, on-device AI features.
npx skills add ml-explore/mlx-swift-lm --all -g -y
Available:

Tells the AI agent to read the bundled SKILL.md before using the API, explain every code change it makes, and never create PRs or post comments on the user's behalf.

All Skills in This Repository (1)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install mlx-swift-lm?โ–ผ

Run `npx skills add ml-explore/mlx-swift-lm --all -g -y` in your terminal to install the skill globally for your AI agent.

How to run LLMs locally on Apple Silicon with Swift?โ–ผ

Add the mlx-swift-lm Swift package to your project, load a model with LLMModelFactory, and chat through the ChatSession API. Models run fully on-device using Apple's MLX framework.

Can mlx-swift-lm run vision-language models?โ–ผ

Yes. The MLXVLM library supports models like Qwen-VL and Gemma that accept images and video alongside text prompts.

Does mlx-swift-lm support fine-tuning and tool calling?โ–ผ

Yes. It includes LoRA and DoRA adapter training for fine-tuning, plus built-in tool calling so models can invoke your Swift functions.

Does mlx-swift-lm work with Claude Code and other AI agents?โ–ผ

Yes. The bundled SKILL.md follows the universal skill standard, so agents like Claude Code, Cursor, and Codex can use it to write correct MLX Swift code.

Related Repositories in Software Engineering

View All in Software Engineeringโ†’