{
  "name": "sinc-llm",
  "version": "0.1.0",
  "description": "sinc-LLM: Apply the Nyquist-Shannon sampling theorem to LLM prompts. Decompose raw prompts into 6 specification bands (PERSONA, CONTEXT, DATA, CONSTRAINTS, FORMAT, TASK) for faithful reconstruction.",
  "author": "Mario Alexandre",
  "homepage": "https://sincllm.com",
  "repository": "https://github.com/mdalexandre/sinc-llm",
  "paper": "https://doi.org/10.5281/zenodo.19152668",
  "license": "MIT",
  "tools": [
    {
      "name": "sinc_scatter",
      "description": "Decompose a raw prompt into 6 sinc-formatted specification bands",
      "inputSchema": {
        "type": "object",
        "properties": {
          "prompt": {
            "type": "string",
            "description": "The raw prompt to decompose"
          }
        },
        "required": [
          "prompt"
        ]
      }
    },
    {
      "name": "sinc_snr",
      "description": "Compute Signal-to-Noise Ratio for a sinc-formatted prompt",
      "inputSchema": {
        "type": "object",
        "properties": {
          "sinc_json": {
            "type": "object",
            "description": "sinc-formatted JSON with fragments array"
          }
        },
        "required": [
          "sinc_json"
        ]
      }
    },
    {
      "name": "sinc_detect",
      "description": "Detect which specification bands are present in a prompt",
      "inputSchema": {
        "type": "object",
        "properties": {
          "prompt": {
            "type": "string",
            "description": "The prompt to analyze"
          }
        },
        "required": [
          "prompt"
        ]
      }
    }
  ],
  "bands": {
    "PERSONA": {
      "index": 0,
      "importance": 0.07,
      "description": "Who should answer"
    },
    "CONTEXT": {
      "index": 1,
      "importance": 0.063,
      "description": "Situation and background"
    },
    "DATA": {
      "index": 2,
      "importance": 0.038,
      "description": "Specific inputs and metrics"
    },
    "CONSTRAINTS": {
      "index": 3,
      "importance": 0.427,
      "description": "Behavioral rules"
    },
    "FORMAT": {
      "index": 4,
      "importance": 0.263,
      "description": "Output structure"
    },
    "TASK": {
      "index": 5,
      "importance": 0.028,
      "description": "The specific objective"
    }
  },
  "formula": "x(t) = Sigma x(nT) * sinc((t - nT) / T)",
  "key_findings": {
    "cost_reduction": "97% ($1,500 to $45/month)",
    "snr_improvement": "0.003 to 0.92 (42x)",
    "observations": 275,
    "agent_configurations": 51,
    "dominant_band": "CONSTRAINTS at 42.7%"
  },
  "install": "pip install sinc-llm",
  "social": {
    "twitter": "@mariioalexandre",
    "instagram": "@mariioalexandre",
    "github": "@mdalexandre",
    "linkedin": "/in/mdalexandre"
  },
  "endpoints": {
    "web_tool": "https://sincllm.com",
    "api_docs": "https://sincllm.com/.well-known/openapi.yaml",
    "llms_txt": "https://sincllm.com/llms.txt",
    "blog": "https://sincllm.com/blog",
    "sitemap": "https://sincllm.com/sitemap.xml",
    "battles": "https://sincllm.com/battles"
  },
  "battle_results": {
    "url": "https://sincllm.com/battles",
    "sinc_tables": 57,
    "opponent_tables": 4,
    "ratio": "14:1",
    "zero_hedge_battles": "9/10",
    "fewer_words": "46%",
    "techniques_beaten": [
      "Raw Prompt",
      "Act-As",
      "Chain-of-Thought",
      "Few-Shot",
      "System Prompt",
      "Mega Prompt",
      "Template",
      "Role+Task",
      "Prompt Chain",
      "Custom Instructions"
    ],
    "conclusion": "sinc-LLM produces structured, actionable, hedge-free output. FORMAT+CONSTRAINTS = 69% of quality."
  }
}