RADAR Documentation

Risk assessment infrastructure for AI agents. Everything you need to integrate RADAR into your agent pipeline.

Installation

npm install @essentianlabs/radar

RADAR runs as a local package in your Node.js project. T1 and T2 assessments execute entirely on your machine — no network calls, no data sent anywhere.

Configuration

Call radar.configure() once at startup. All fields are optional — RADAR works locally with zero config, but an API key unlocks enhanced suggestions and T3/T4 capabilities.

import { radar } from '@essentianlabs/radar';

radar.configure({
  apiKey:  'rdr_live_xxxx',    // optional — enables enhanced suggestions + T3/T4
  llmKey:  'sk-ant-xxxx',      // your own LLM key for local T1/T2 scoring
  activities: {
    email:      'conservative',   // permissive | balanced | conservative
    financial:  'conservative',
    publishing: 'balanced',
    deletion:   'conservative'
  },
  webhookUrl: 'https://your-app.com/radar-escalation'  // T4 escalation target (paid)
});

Configuration fields

Your first assessment

Wrap any output action with radar.assess(). It returns a verdict telling your agent whether to proceed.

const result = await radar.assess(
  'Send price increase email to 50,000 users',
  'email'
);

console.log(result);
// {
//   proceed: false,
//   tier: "T3",
//   risk_score: 14,
//   suggestion: {
//     recommended: "mitigate",
//     options: {
//       proceed: "Send as-is to all 50,000 users",
//       mitigate: "Send in batches under 1,000 with a 24h delay between batches",
//       block: "Do not send — requires manager approval for bulk pricing changes",
//       escalate: "Route to human decision-maker"
//     }
//   }
// }

Response fields

The four decision options are always: proceed, mitigate, block, and escalate. Your agent can follow the recommended option or apply its own logic.

POST /radar/assess

Submit an action for server-side risk assessment. Returns a tier assignment, risk score, and suggestions. Used for T3/T4 when the local SDK escalates to the server panel.

Request

Example request

curl -X POST https://radar.essentianlabs.com/radar/assess \
  -H "Authorization: Bearer rdr_live_xxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "Delete all user records older than 2 years",
    "activity_type": "deletion",
    "agent_id": "data-agent"
  }'

Example response

{
  "tier": "T4",
  "risk_score": 21,
  "proceed": false,
  "suggestion": {
    "recommended": "escalate",
    "options": {
      "proceed": "Delete all matching records immediately",
      "mitigate": "Archive records instead of deleting; purge after 30-day review",
      "block": "Do not delete — requires DBA and compliance sign-off",
      "escalate": "Route to database administrator for manual review"
    }
  },
  "escalation_id": "esc_abc123"
}

Error codes

POST /radar/suggest

Request enhanced suggestions from the EssentianLabs server. Only anonymised signals are sent — no action text is transmitted.

Request

Example response

{
  "enhanced_suggestion": "Based on aggregate patterns for this activity type, consider adding a rate limit before proceeding.",
  "confidence": 0.82
}

POST /radar/register

Register for a free API key. No credit card required. Enables enhanced suggestions immediately.

Request

Example

curl -X POST https://radar.essentianlabs.com/radar/register \
  -H "Content-Type: application/json" \
  -d '{ "email": "dev@example.com" }'

{
  "api_key": "rdr_live_abc123def456",
  "plan": "free",
  "message": "API key created. Store it securely — it won't be shown again."
}

POST /radar/config

Update your activity slider positions server-side. Accepts permissive, balanced, or conservative. Synced to the local SDK on next assessment.

Request

{
  "activities": {
    "email": "conservative",
    "financial": "conservative",
    "publishing": "balanced"
  }
}

POST /radar/webhook

Register or update your T4 escalation webhook URL. When an action reaches T4, RADAR sends a POST to this URL with the escalation payload.

Webhook payload (sent to your URL)

{
  "escalation_id": "esc_abc123",
  "tier": "T4",
  "risk_score": 22,
  "activity_type": "financial",
  "suggestion": { /* same shape as /assess response */ },
  "timestamp": "2026-03-26T14:30:00Z"
}

GET /radar/history

Returns the last 100 scoring metadata records. No action content is stored or returned — only anonymised signals.

Example response

{
  "records": [
    {
      "id": "rec_001",
      "tier": "T2",
      "risk_score": 7,
      "activity_type": "email",
      "verdict": "mitigate",
      "timestamp": "2026-03-26T10:15:00Z"
    }
  ],
  "total": 47
}

GET /radar/health

Health check endpoint. Returns server status.

{
  "status": "ok",
  "version": "1.0.0"
}

Tier logic

Every action is scored from 1 to 25. The score maps to a tier based on the slider value for that activity type.

Score ranges above are defaults at the Balanced slider position. The slider shifts these thresholds:

Slider configuration

Each activity type has a slider with three positions: Permissive, Balanced, and Conservative. The slider controls where the tier boundaries fall.

At conservative settings, more actions escalate to higher tiers. At permissive settings, only the riskiest actions leave T1.

Local risk register

RADAR maintains a local SQLite database at .radar/register.db in your project directory. This register:

• Stores action hashes (SHA-256), never action text
• Feeds context into future assessments (repeat detection, pattern analysis)
• Is entirely under your control — delete it, back it up, or inspect it anytime
• Never synced to any server

Privacy model

T1 and T2 are fully local. No action content ever reaches EssentianLabs servers. Scoring happens on your machine using your own LLM key.

T3 and T4 transmit action text to the EssentianLabs server panel for multi-model deliberation. This text is processed in memory and immediately discarded — never written to disk, never stored in any database, never used for training.

What we store (all plans): anonymised scoring metadata only.

Nothing in this schema can identify your users or reveal what your agent was doing.

Pricing

Free tier includes T1/T2 local assessment, local risk register, and enhanced suggestions with an API key. No limits.

Paid tier adds T3 multi-model deliberation, T4 human escalation via webhook, hosted risk register, and the risk register dashboard.

Early access pricing — contact hello@essentianlabs.com