CR Gateway

How It Works

Three steps. Zero complexity.

Keep your LLM provider, your keys, your stack. CR Gateway sits between your agent and its next action.

Your Agent Calls Its LLM

Your agent makes LLM calls as usual. Your provider, your API key, your cost. Nothing changes.

POST to CR Gateway

Send the LLM response to CR Gateway for validation: safety checks, confidence scoring, hallucination detection.

Get Validation Results

Receive valid/invalid verdict, scores, and recommendations. Bad chains killed before they waste more LLM calls.

Features

Everything your agents need

Six capabilities, one API. No SDKs required — just HTTP.

⚡

Validation

Safety, confidence, hallucination, danger term detection. Every response checked before your agent acts on it.

Sub-millisecond

☠

Swarm Fail-Fast

Kill bad agent chains early using chain-rule probability. One weak link flags the whole chain before it burns tokens.

Save 20-40% on LLM calls

📦

Context Compression

10 messages compressed into 1 briefing. Pure CPU, no LLM call needed. Feed your agent less, get the same quality.

44%+ token reduction

🗃

Message Storage

VERNOT-encoded messages stored 33% smaller than JSON. Built-in audit trail for compliance and debugging.

33% smaller storage

🔗

Agent Relay

Validated agent-to-agent messaging. JSON in, JSON out. Every relay message validated before delivery.

JSON in / JSON out

🛡

Hallucination Guard

Catches fabricated URLs, invented data, overconfident claims. Five independent checks on every message.

5-check pipeline

Quick Start

One request. Instant validation.

No SDK needed. Standard HTTP from any language.

curl -X POST https://cr-gateway-worker.jnowlan21.workers.dev/v1/validate \
  -H "Content-Type: application/json" \
  -H "X-API-Key: YOUR_KEY" \
  -d '{
    "message": {
      "type": "analysis",
      "content": "Based on market data...",
      "confidence": 0.85
    }
  }'

const resp = await fetch('https://cr-gateway-worker.jnowlan21.workers.dev/v1/validate', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
    'X-API-Key': 'YOUR_KEY'
  },
  body: JSON.stringify({
    message: {
      type: 'analysis',
      content: 'Based on market data...',
      confidence: 0.85
    }
  })
});

const result = await resp.json();
// result.valid === true
// result.checks.hallucination.passed === true

import requests

resp = requests.post(
    'https://cr-gateway-worker.jnowlan21.workers.dev/v1/validate',
    headers={
        'Content-Type': 'application/json',
        'X-API-Key': 'YOUR_KEY'
    },
    json={
        'message': {
            'type': 'analysis',
            'content': 'Based on market data...',
            'confidence': 0.85
        }
    }
)

result = resp.json()
# result['valid'] == True
# result['checks']['hallucination']['passed'] == True

API Reference

Endpoints

All endpoints accept and return JSON. Authentication via X-API-Key header.

Method	Path	Description
POST	/v1/validate	Validate a message (safety, confidence, hallucination)
POST	/v1/store	Validate + store a message with VERNOT encoding
POST	/v1/compress	Compress conversation context (CPU-only, no LLM)
POST	/v1/swarm/check	Fail-fast chain check for agent swarms
POST	/v1/relay	Validated agent-to-agent relay
POST	/v1/onboard	Self-service API key provisioning
GET	/health	Health check

Pricing

Start free. Scale when ready.

No credit card required. Upgrade anytime.

Free

Free forever

1,000 requests / day
2 guard checks per request
30-day message storage
Community support

Pro

$49/mo

For production workloads

500,000 requests / month
All validation checks
90-day message storage
Configuration API
Email support

Scale

$199/mo

For high-volume teams

5,000,000 requests / month
All validation checks
365-day message storage
Replay & export
Priority support

Validation infrastructure for AI agents