Model Library
Every local model, version and adapter in one organized place.
v1.1.0 · July 2026 · Apple Silicon only
MLX Forge
The complete AI laboratory for Apple Silicon. The entire model lifecycle — from Hugging Face discovery to a production endpoint — in one native app on Apple's MLX framework.
One pipeline, nine stages
Click a stage — every one is a first-class studio in the app.
No-code at the start, full code at the end
The one non-negotiable principle: every visual action can emit the equivalent reproducible code — Python mlx-lm, Swift MLX, CLI, or curl.
🏛️
🧪
Playground
Chat & completion sandbox with full parameter control.
📚
Datasets
Import, inspect and split training data for fine-tuning.
📓
Notebooks
Scriptable experiments living next to your models.
🔩
Metal Lab
GPU-level profiling for the performance-obsessed.
👁️
Vision Lab
Multimodal experimentation with vision models.
🖥️
Cluster
Fan work out to other Macs on your network (Bonjour).
⌘K
Command palette
Everything in the app is one keystroke away.
Point any OpenAI SDK at your Mac
ForgeServer speaks the real OpenAI wire format — verified with golden tests against the unmodified openai Python SDK.
python
from openai import OpenAI
client = OpenAI(base_url="http://localhost:8080/v1", api_key="mlxf-…")
resp = client.chat.completions.create(
model="qwen3-14b-4bit",
messages=[{"role": "user", "content": "Hello from my own silicon"}],
stream=True,
)Forge something
Free download. Your models, your data, your silicon.
⬇ Download MLX Forge 1.1.0