How can we help?

Before contacting support

1. Open Settings and refresh the hardware reading.
2. Confirm the model is registered and supports the selected pipeline.
3. Check that the model library has the required free space plus safety headroom.
4. For MiniMax H3, select a current sd-cli build in Settings and use 24 FPS with a 17n+5 frame count.
5. For custom models, confirm the exact Diffusers pipeline class and immutable Hugging Face commit.
6. Export redacted diagnostics from Settings and attach the file if comfortable.

What to include

Include the app version, macOS version, Mac chip, unified memory, model name, runtime, selected pipeline, generation size, frame count, and exact error. Never include Hugging Face tokens, local API tokens, private source media, or proprietary model files.

Common questions

Does generation use BytesWeavers servers?

No. Rendering is local. Network access is used for downloads, StoreKit, links you open, and features you explicitly enable.

Why is a model marked incompatible?

The model’s declared minimum unified memory or free-storage requirement exceeds the current Mac profile. Closing apps can improve available memory but cannot change total unified memory.

Can every Hugging Face model work?

No universal runtime can execute every architecture. Custom support works when a repository matches one of the app’s maintained adapters and the selected pipeline class exposes the required inputs.

How do I cancel Pro?

Use Manage Subscription inside the App or your App Store account settings. Apple handles billing and cancellation.

Contact

Email bytesweavers@gmail.com. Typical response target: two business days. Support cannot provide legal advice about model or output licensing.