Strands MLX: Apple Silicon Model Provider 🍎

Run Strands Agents locally on Apple Silicon with inference, fine-tuning, and vision.

MLX provider for Strands Agents with LoRA training pipelines.

Installation

# Requires macOS with Apple Silicon (M1/M2/M3/M4)
pip install strands-mlx strands-agents-tools

Quick Start

from strands import Agent
from strands_mlx import MLXModel
from strands_tools import calculator

model = MLXModel(model_id="mlx-community/Qwen3-1.7B-4bit")
agent = Agent(model=model, tools=[calculator])

agent("What is 29 * 42?")

Architecture

graph LR
    A[Agent Conversations] -->|MLXSessionManager| B[Training Data JSONL]
    B -->|dataset_splitter| C[train/valid/test]
    C -->|mlx_trainer| D[LoRA Adapter]
    D -->|MLXModel| E[Domain Expert Agent]
    E -.->|Continuous Learning| A

    style A fill:#e1f5ff
    style E fill:#d4edda
    style D fill:#fff3cd

The complete training cycle: Agents collect their own training data → fine-tune themselves → become domain experts → continue learning.

---

Train Your Own Model

1. Collect Training Data

from strands import Agent
from strands_tools import calculator
from strands_mlx import MLXModel, MLXSessionManager, dataset_splitter, mlx_trainer

agent = Agent(
    model=MLXModel(model_id="mlx-community/Qwen3-1.7B-4bit"),
    session_manager=MLXSessionManager(session_id="my_training", storage_dir="./dataset"),
    tools=[calculator, dataset_splitter, mlx_trainer],
)

# Have conversations - auto-saved to JSONL
agent("Teach me about quantum computing")
agent("Calculate 15 * 7")

# Saved to: ./dataset/my_training.jsonl

2. Split Dataset

agent.tool.dataset_splitter(
    input_path="./dataset/my_training.jsonl"
)
# Creates train.jsonl, valid.jsonl, test.jsonl (80/10/10 split)

3. Train with LoRA

agent.tool.mlx_trainer(
    action="train",
    config={
        "model": "mlx-community/Qwen3-1.7B-4bit",
        "data": "./dataset/my_training",
        "adapter_path": "./adapter",
        "iters": 200,
        "learning_rate": 1e-5,
        "batch_size": 1
    }
)

4. Use Trained Model

from strands import Agent
from strands_mlx import MLXModel

trained = MLXModel("mlx-community/Qwen3-1.7B-4bit", adapter_path="./adapter")
agent = Agent(model=trained)

agent("Explain quantum computing")  # Uses trained knowledge!

---

Vision Models

from strands_mlx import MLXVisionModel

model = MLXVisionModel(model_id="mlx-community/Qwen2-VL-2B-Instruct-4bit")
agent = Agent(model=model)

agent("Describe: <image>photo.jpg</image>")
agent("Transcribe: <audio>speech.wav</audio>")
agent("What happens: <video>clip.mp4</video>")

---

Training Tools

Tool	Purpose
mlx_trainer	Background LoRA training
dataset_splitter	Split JSONL → train/valid/test
validate_training_data	Check format & token counts
mlx_invoke	Runtime model switching

---

Recommended Models

Model	Size	Use Case
mlx-community/Qwen3-1.7B-4bit	~1GB	Fast local inference
mlx-community/Qwen3-8B-4bit	~4GB	Balanced performance
mlx-community/Qwen2-VL-2B-Instruct-4bit	~2GB	Vision tasks

---

Why MLX?

• Native Apple Silicon: Optimized for M1/M2/M3/M4 chips
• No GPU drivers: Uses Metal directly
• Fast inference: Comparable to server GPUs
• Local privacy: No data leaves your machine
• Fine-tuning: LoRA training on consumer hardware

---

Train your own domain expert agent—right on your Mac. 🍎