{
  "_instructions": {
    "purpose": "Single source of truth for the local MeltMind language-model runtime.",
    "note": "Runtime code should read this file instead of hard-coding model identifiers or generation settings."
  },
  "provider": "llama.cpp",
  "model": {
    "family": "MiniCPM4",
    "variant": "MiniCPM4-8B",
    "parameter_count_billions": 8,
    "openbmb_challenge_eligible": true,
    "tiny_titan_eligible": false,
    "hugging_face_repo": "openbmb/MiniCPM4-8B",
    "gguf_repo": "AyyYOO/MiniCPM4-8B-Q4_K_M-GGUF",
    "gguf_file": "minicpm4-8b-q4_k_m.gguf",
    "chat_template": "chatml"
  },
  "server": {
    "base_url": "http://127.0.0.1:8080",
    "chat_completions_path": "/v1/chat/completions",
    "health_path": "/health",
    "request_timeout_seconds": 120
  },
  "generation": {
    "temperature": 0.35,
    "top_p": 0.9,
    "max_tokens_chat": 280,
    "max_tokens_designer_explanation": 500,
    "stream": false
  },
  "context": {
    "maximum_retrieved_chunks": 3,
    "maximum_conversation_turns": 8,
    "include_structured_tool_results": true,
    "include_uncertainty_instructions": true
  }
}