sinc-LLM Scatter API Documentation

Your API key was sent to the email address associated with your Stripe purchase. Include it in the Authorization header as a Bearer token.

Authorization: Bearer YOUR_API_KEY

If you did not receive your API key, contact mdalexandre@duarteluxury.digital with your Stripe receipt.

POST /api/scatter

Decomposes a raw text prompt into a 6-band sinc JSON object with PERSONA, CONTEXT, DATA, CONSTRAINTS, FORMAT, and TASK fragments.

Request

{
  "prompt": "Write a Python script that scrapes product prices from Amazon and saves them to a CSV file"
}

Response (200 OK)

{
  "formula": "x(t) = \u03a3 x(nT) \u00b7 sinc((t - nT) / T)",
  "T": "specification-axis",
  "fragments": [
    {"n": 0, "t": "PERSONA", "x": "Senior Python developer with web scraping expertise..."},
    {"n": 1, "t": "CONTEXT", "x": "Production scraping task targeting Amazon product pages..."},
    {"n": 2, "t": "DATA", "x": "Input: Amazon product URLs or search queries..."},
    {"n": 3, "t": "CONSTRAINTS", "x": "MUST use requests + BeautifulSoup. MUST handle rate limiting. MUST respect robots.txt. NEVER store personal data. MUST handle pagination. MUST validate prices as numeric. MUST include user-agent rotation. MUST retry on 429/503 with exponential backoff..."},
    {"n": 4, "t": "FORMAT", "x": "Output: Python script with CSV output, requirements.txt, usage instructions..."},
    {"n": 5, "t": "TASK", "x": "Write a complete, runnable Amazon price scraper with CSV export."}
  ]
}

Error Responses

GET /api/scatter/health

Health check endpoint. No authentication required.

Response (200 OK)

{"status": "ok", "version": "1.0.0", "model": "sinc-scatter-v1"}

Rate limit headers are included in every response: X-RateLimit-Remaining, X-RateLimit-Reset.

Python

import requests

API_KEY = "your_api_key_here"
URL = "https://sincllm.com/api/scatter"

response = requests.post(
    URL,
    json={"prompt": "Build a REST API for a todo app with authentication"},
    headers={"Authorization": f"Bearer {API_KEY}"}
)

sinc_json = response.json()
print(f"Generated {len(sinc_json['fragments'])} bands")

# Extract the CONSTRAINTS band (the largest, ~42.7% of content)
constraints = next(f for f in sinc_json["fragments"] if f["t"] == "CONSTRAINTS")
print(f"Constraints: {constraints['x'][:200]}...")

JavaScript (Node.js / Browser fetch)

const API_KEY = "your_api_key_here";

const response = await fetch("https://sincllm.com/api/scatter", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    "Authorization": `Bearer ${API_KEY}`
  },
  body: JSON.stringify({
    prompt: "Build a REST API for a todo app with authentication"
  })
});

const sincJson = await response.json();
console.log(`Generated ${sincJson.fragments.length} bands`);

// Pass the sinc JSON directly to your LLM API
const llmPrompt = JSON.stringify(sincJson);
// Use llmPrompt as the system message or user message to Claude/GPT

curl

curl -X POST https://sincllm.com/api/scatter \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{"prompt": "Build a REST API for a todo app with authentication"}'

The recommended integration pattern is to use the Scatter API as a preprocessing step before sending prompts to any LLM:

Raw Prompt
    |
    v
[POST /api/scatter]  -->  sinc JSON (6-band decomposition)
    |
    v
[Your LLM API]  -->  Higher fidelity response
    |                  (reduced hallucination,
    v                   better constraint adherence)
Your Application

The sinc JSON output can be passed directly as the prompt to Claude, GPT-4, Gemini, Llama, or any instruction-following LLM. The structured format improves response quality across all models.