InferrLM adds Apple Silicon MLX support

InferrLM is quietly turning into infrastructure, not just another chat UI. MLX support matters because it adds a native Apple Silicon path and makes the local-AI stack feel more serious for power users.

• –Backend flexibility is the real win here: MLX only helps the Apple Silicon slice, but that is the group most sensitive to native performance and battery efficiency.
• –The app already bundles local inference, RAG, OCR, camera input, and a network server, so this release strengthens an already coherent workflow.
• –The first Reddit reply is already asking about specific MLX model compatibility, which is the practical question local-AI users care about.
• –Open-source status plus an AGPL license keep the implementation inspectable and forkable for people who want to tinker.