Sentinel Signal Systems

Signal Before Action.

Infrastructure-grade scoring for automated decision workflows.

A scoring layer for agents that need to decide what to do next.

Deterministic • Calibrated • Versioned • Explainable

Get an API key Read the docs Interactive tools

What it is

Sentinel Signal provides a simple scoring primitive you can drop into automated workflows and agents.

Send a payload -> get a stable score, confidence, and explanation back.

Close the loop with POST /v1/feedback using scoring x-request-id values.

No dashboards. No fluff. Just signal.

Quickstart (copy/paste)

Request

curl -s https://sentinelsignal.io/v1/score \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "workflow": "healthcare.denial",
    "payload": {
      "external_claim_id": "CLM-SS-0001",
      "service_date": "2026-02-14",
      "provider_id": 1021,
      "payer_id": 44,
      "cpt_code": "99213",
      "icd10_code": "E119",
      "billed_amount": 250.00,
      "place_of_service": "11",
      "network_status": "in_network"
    },
    "options": {
      "operating_point": "high_recall",
      "distribution_profile": "commercial_beta"
    }
  }'

Response

{
  "workflow": "healthcare.denial",
  "score": 0.81,
  "risk_level": "high",
  "confidence": 0.92,
  "model_version": "0.3.3",
  "explanation": {
    "top_factors": [
      "High billed amount vs typical band",
      "CPT/payer prior denial pattern",
      "Out-of-policy place-of-service mix"
    ]
  },
  "details": {
    "workflow": "healthcare.denial",
    "denial": {
      "denial_probability": 0.81,
      "operating_point": "high_recall",
      "distribution_profile": "commercial_beta",
      "routing_source": "explicit"
    }
  },
  "scored_at": "2026-02-14T12:00:00Z"
}

OpenAPI: /openapi.json | Docs: /portal/docs-home

Workflows (v1)

healthcare.denial

Risk score for claim denial likelihood with calibrated confidence and factor explanations.

Why agents use it

  • Deterministic outputs: identical inputs -> identical outputs (per model version)
  • Versioned responses: every score includes model_version
  • Explainable by default: top factors included for downstream reasoning
  • Regime-aware denial routing metadata included in response details
  • Built for automation: stable schema, low latency, simple auth

Portal Tools

The main portal now includes the newer staging validation tools from Portal Lite: self-serve dashboard signup/login + API key management, interactive Swagger/ReDoc, and the claim score + appeal demo flow.

Use /portal/dashboard to sign up, log in, and create/rotate API keys.

Interactive docs and demo requests run through the staging browser proxy at /portal-lite/staging.

Dashboard Swagger UI ReDoc Demo Page

Pricing

Free: 1,000 calls/month

Usage: $0.003 per scoring call after the first 1,000 monthly calls (Stripe metered billing)

Enterprise: volume + SLA

Checkout + key lifecycle APIs are available in the console and docs.

View pricing Start free

Trust & governance

Scores are decision support, not decisions.

No payload retention by default (processed in-memory).

Feedback notes are short, structured, and screened for PHI-like patterns.

Observability endpoints: /metrics is public Prometheus scrape output, while /metrics.json is a protected snapshot (bearer token with metrics:read, schema 1).

Latest validation snapshot (2026-02-17 UTC): local full 5-seed run p95 latency 66.87 ms; Fly dual-machine rerun p95 latency 529.02 ms on API profile 2 x shared 4 vCPU / 4096 MB.

Drift-aware scoring and model governance processes available in docs.

Model validation & governance | Security overview