Document speech-swift MLX engine usage

README.md

@@ -96,16 +96,23 @@ path (STFT → features → ERB mask + deep filtering → iSTFT), documented in
 [upstream repository](https://github.com/Rikorose/DeepFilterNet).
 
 On Apple devices, [speech-swift](https://github.com/soniqo/speech-swift)
-provides ready-made speech enhancement via the
-[CoreML variant](https://huggingface.co/aufklarer/DeepFilterNet3-CoreML):
+runs this model directly via its MLX engine (or use the
+[CoreML variant](https://huggingface.co/aufklarer/DeepFilterNet3-CoreML) for
+the Neural Engine):
 
 ```swift
 import SpeechEnhancement
 
-let enhancer = try await SpeechEnhancer.fromPretrained()
+let enhancer = try await SpeechEnhancer.fromPretrained(engine: .mlx)
 let clean = try enhancer.enhance(audio: noisyAudio, sampleRate: 48000)
 ```
 
+CLI:
+
+```bash
+swift run speech denoise noisy.wav --engine mlx
+```
+
 ## Tensor contract
 
 Channels-last `[B, T, F, C]` — time is the conv H axis, frequency the W axis.