Support
We're here to help.
Contact
For questions, bug reports, or feedback, email us at 254553782@qq.com. We aim to respond as soon as possible.
System Requirements
- Mac with Apple Silicon (M1 or later)
- macOS 14 Sonoma or later
- Enough free disk space to store the model you download (several gigabytes per model)
- Sufficient unified memory to load the selected model
Frequently Asked Questions
Does sMLX work offline?
Yes. Once a model is downloaded, all inference runs entirely on your Mac with no internet connection required. The only time the App needs the network is when you download a model.
Where are my conversations stored?
They aren't. Conversations are processed in memory and are not persisted to disk by the App. They never leave your device and are gone when you close the session.
Why is the first download large or slow?
Models are downloaded directly from Hugging Face and can be several gigabytes in size. Download speed depends on your internet connection. Subsequent launches reuse the downloaded model and start quickly.
A model download failed. What should I do?
Check your internet connection and available disk space, then try again. If a download was interrupted, deleting the partial files and re-downloading usually resolves the issue. If problems persist, contact support.
Which models are available, and what are their licenses?
The App includes Qwen 3.6-27B (MTP) and Qwen 3-8B, both under Apache 2.0. See the Licenses page for full details and attribution.
Is the model output reliable?
AI-generated content can be inaccurate or incomplete. Always verify important information independently. See our Terms of Use for the full disclaimer.