{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreihkr5mwwcfooa2zq7fyq4mx7yuuhf5p5l4pq3gvldxf42yiexqdk4",
    "uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mhcqp2tvyev2"
  },
  "path": "/t/m-courtyard-v0-5-0-lm-studio-ollama-as-parallel-local-runtime-options-on-apple-silicon/174367#post_1",
  "publishedAt": "2026-03-18T02:44:12.000Z",
  "site": "https://discuss.huggingface.co",
  "textContent": "M-Courtyard v0.5.0 is now live.\n\nM-Courtyard is a native macOS desktop app for local LLM fine-tuning on Apple Silicon, built around Apple MLX. It covers the full workflow from document import and dataset generation to LoRA fine-tuning, testing, and export.\n\nWhat’s new in v0.5.0:\n\n  * LM Studio and Ollama are now presented as parallel local runtime options for dataset generation\n  * Better LM Studio connection guidance and status feedback inside the app\n  * Clearer support positioning across the product and docs\n  * Export support for Ollama, MLX, and GGUF\n  * Built-in rules remain available when users want a no-runtime dataset path\n\n\n\nThis release is mainly about making the local workflow more accurate, practical, and easier to understand.\n\nI’m attaching a few updated screenshots below. Feedback is welcome.\n\n\n\n\n\n",
  "title": "M-Courtyard v0.5.0: LM Studio + Ollama as Parallel Local Runtime Options on Apple Silicon"
}