mlx 2
Native Rapid-MLX provider for DeepSeek Harness β teaches DSH what the local server already knows (memory-fitted context via max_model_len), plus model-management tools and a /rapid-mlx overview.
dsh plugin add @raullenchai/dsh-providerLocal model manager plugin for DeepSeek Harness: discover models, one-click configure, start/stop oMLX and llama.cpp, monitor memory and throughput, register into DSH.
dsh plugin add dsh-local-model-manager