SENTINEL SHIELD

Shield Your Platform From Abuse

AI-powered Trust & Safety that stops spam, fraud, and abuse in real-time. Built on the same engine that protects Discord's 200M+ users.

View Documentation
<50ms
Verdict Latency
99.9%
Target availability
316x
Faster Than Manual Review

How It Works

Three steps from zero to full Trust & Safety coverage.

01

Connect

Install our SDK and send events with 3 lines of code.

pip install liquidcortex-shield

from liquidcortex_shield import Shield
shield = Shield(api_key="lc_shield_...")
02

Protect

Rules evaluate every event in real-time. Threats are auto-blocked before they reach your users.

verdict = await shield.evaluate(
  event_type="user_message",
  data={"user_id": "u_8291", ...}
)
# verdict.action → "allow" | "reject"
03

Investigate

Dashboard shows entity profiles, risk scores, and a full audit trail for every decision.

# Dashboard: /dashboard/shield
# Entity profiles with risk scores
# Full audit trail for every verdict
# Exportable compliance reports

Everything You Need

A complete Trust & Safety toolkit, not just a classifier. Shield protects against Shadow AI — unauthorized AI tools accessing your company data without IT oversight. Every AI action is logged, permissioned, and auditable.

Real-Time Verdicts

Sub-50ms synchronous API for inline blocking. Evaluate events before they reach your users with zero perceived latency.

AI-Generated Rules

Describe rules in plain English and our AI writes production-ready SML. No regex, no code, just intent.

20+ Rule Templates

Deploy spam filters, fraud detection, bot blocking, and NSFW classifiers instantly from our template library.

Entity Intelligence

Track users, IPs, emails, and devices across every event. Apply labels, build risk profiles, and correlate threats.

Compliance Built-In

DSA Article 15 and UK Online Safety Act reporting with one click. Transparency reports generated automatically.

Webhook Actions

POST verdicts to your systems in real-time. Every payload HMAC-SHA256 signed for tamper-proof delivery.

5 Minutes to Integration

A single API call stands between your platform and real-time abuse protection.

shield_example.py
from liquidcortex_shield import Shield

shield = Shield(api_key="lc_shield_...")

# Real-time verdict on user message
verdict = await shield.evaluate(
    event_type="user_message",
    data={
        "user_id": "u_8291",
        "content": "Buy followers cheap! Visit spam.site",
        "ip": "203.0.113.42",
        "channel": "general"
    }}
)

if verdict.action == "reject":
    return {"error": "Message blocked by safety policy"}

Shield Pricing

Separate from your AI Employee subscription. Start free, scale as you grow.

SHIELD STARTER
$499/mo

For growing platforms that need basic protection.

  • 500K events/mo
  • 10 active rules
  • Email alerts
  • Basic dashboard
  • Community support
  • 7-day event retention
RECOMMENDED
SHIELD PRO
$2,499/mo

For platforms with serious Trust & Safety needs.

  • 5M events/mo
  • Unlimited rules
  • Webhooks + Investigation UI
  • AI rule review (weekly)
  • Priority support
  • 90-day event retention
SHIELD ENTERPRISE
$9,999+/mo

For regulated industries and high-scale platforms.

  • Unlimited events
  • Unlimited rules + custom UDFs
  • Dedicated instance + availability target
  • DSA/OSA compliance reports
  • Dedicated CSM
  • Target: 99.9% availability (post-launch)
Contact Sales

BUILT ON PROVEN INFRASTRUCTURE

Powered by ROOST's Osprey — the same engine protecting Discord, Bluesky, and Matrix.org

Discord
Bluesky
Matrix.org
Mastodon

Frequently Asked Questions

Shield protects against Shadow AI — unauthorized AI tools accessing your company data without IT oversight. With Shield, every AI action is logged, permissioned, and auditable. Instead of employees using random AI tools with no controls, Shield ensures all AI activity flows through a governed pipeline with full visibility.

OpenAI's moderation API is a single-purpose classifier. Shield is a full Trust & Safety platform with customizable rules, entity tracking across sessions, investigation tools, and compliance reporting. You define the policies, not a black-box model.

Yes. Shield is billed independently from your AI Employee subscription. It works alongside your existing LiquidCortex plan and can even be used as a standalone product without an AI Employee subscription.

Most teams are live in under 5 minutes with our Python SDK. Install the package, initialize with your API key, and call shield.evaluate() on your event stream. We also offer REST endpoints for non-Python stacks.

Shield generates transparency reports compliant with DSA Article 15 (EU Digital Services Act) and the UK Online Safety Act. Reports include moderation volumes, appeal rates, response times, and automated decision breakdowns.

Start Protecting Your Platform Today

14-day free trial, no credit card required.