Sentinel AI
Database & Security Autopilot
Your server fixes itself.
Autonomous database optimization, security monitoring, and AI-powered self-healing for MySQL, PostgreSQL, Redis, and Fail2Ban. Ask your server anything via AI Chat. Sentinel handles the firefighting so you can focus on building.
Key achievements
12+
Production servers
847
Issues fixed this week
60+ days
Uptime streak
Your server gets an AI brain
Enterprise tier adds Claude AI directly on your server. Ask questions, get autonomous healing, and share threat intelligence across your entire fleet — all with your own API key.
AI Chat — Ask your server anything
Real-time streaming chat with Claude AI running directly on your server. No data leaves without your permission — uses your own Anthropic API key.
- ✦Check logs, services, Docker containers in plain English
- ✦Restart services, ban attack subnets, query your DB
- ✦Read config files for diagnostics (/etc/caddy, /etc/nginx…)
- ✦Persistent memory — remembers your server across sessions
- ✦Semantic rules learned from every conversation
AI Healer — 24/7 autonomous self-healing
Every 5 minutes Sentinel checks for problems. When something breaks, Claude decides what to do and executes safe fixes automatically — no alerts, no waiting.
- ✦Restarts crashed services (nginx, mysql, redis…)
- ✦Detects high CPU/RAM and takes action
- ✦Persistent memory learns your server patterns
- ✦Full audit log — every decision explained
- ✦Cooldown system prevents loops
Cross-Fleet IP Reputation
Attackers banned on one server get blocked on all your servers — automatically, before they even knock. No third-party feeds required.
- ✦Fail2ban bans reported to Queen on every heartbeat
- ✦Queen correlates IPs across your entire fleet
- ✦Knowledge Book entry created per attacker IP
- ✦All servers sync and pre-ban known attackers
- ✦Works with any fail2ban jail (ssh, http, mysql…)
Core Autopilot Capabilities
Available on all plans — Sentinel watches your databases, cache, security and system health 24/7.
Security Autopilot
Integrated with Fail2Ban and CrowdSec to detect brute-force attacks, manage jails, and secure SSH/web entry points. Automatically syncs security state with running services. Blocks 95% of attacks before they reach your application.
Database Autopilot
Analyzes MySQL slow logs, runs EXPLAIN on queries, and detects performance pathologies. Automatically applies safe index suggestions during scheduled night windows. Turn 850ms queries into 12ms instantly.
Cache Guardian
Real-time monitoring for Redis and PHP OPCache. Tracks memory usage, hit ratios, and evictions, surfacing bottlenecks before they impact performance. Proactive health checks for your data layer.
DB Doctor & Diagnostics
A dedicated cockpit for deep database inspection. Monitors buffer pools, long transactions, and table fragmentation with safe, read-only diagnostics and manual fix shortcuts.
Queen Integration
Connects to a central intelligence platform for cross-fleet knowledge sharing. Sentinel syncs the Knowledge Book every hour, applying Queen-verified solutions proactively. ROI cost profiles, IP reputation, and AI-analyzed patterns — all managed centrally.
System Health Dashboard
Mission control cockpit with high-level readiness indicators and global counters for actionable, diagnostic, and fixed issues across your entire infrastructure.
Why Sentinel vs Traditional Monitoring
| Feature | Traditional Monitoring | Sentinel AI |
|---|---|---|
| Detects issues | ✓ | ✓ |
| Sends alerts | ✓ | ✓ |
| Auto-fixes issues | ✗ | ✓ |
| Learns & adapts | ✗ | ✓ |
| Database optimization | ✗ | ✓ |
| Security autopilot | ✗ | ✓ |
| AI Chat (BYOK — your API key) | ✗ | ✓ Enterprise |
| AI self-healing (24/7 background) | ✗ | ✓ Enterprise |
| Cross-fleet IP reputation | ✗ | ✓ Enterprise |
| Persistent AI memory (cross-session) | ✗ | ✓ Enterprise |
How it works
Get started in 60 seconds with our lightweight agent.
Install the agent
One-line install script for Ubuntu, Debian, CentOS, RHEL, and Fedora. No complex setup required.
Agent connects
Sentinel securely connects to your server and starts monitoring.
Autopilot activates
Database, cache, and security optimizations run automatically.
Frequently Asked Questions
Is Basic really free?
Yes, Basic is completely free for 1 production server. You get system monitoring, security detection, and manual database optimizations at no cost.
How does the agent work?
The agent is a lightweight Python script that runs on your server. It monitors MySQL/PostgreSQL, Redis, Fail2Ban, and system metrics. All data is encrypted and sent to Queen for analysis.
What happens if something goes wrong?
Sentinel only applies safe, explainable fixes. Every action is logged in the audit trail. You can disable any autopilot feature at any time.
Which operating systems are supported?
Sentinel runs natively on Ubuntu, Debian, CentOS, RHEL, and Fedora. It installs as a systemd service, so no Docker is required on your server.
Can I use Sentinel with my existing stack?
Yes, Sentinel works with MySQL, PostgreSQL, Redis, PHP OPCache, Fail2Ban, and CrowdSec. It's designed to complement your existing infrastructure.
How does AI Chat work — does my data go to Anthropic?
AI Chat uses your own Claude API key (BYOK — Bring Your Own Key). Messages go from your server directly to Anthropic's API. We never see your conversations. The AI Chat feature is only available on Enterprise and requires you to configure your own API key in Sentinel's settings.
What is AI Healer and is it safe?
AI Healer is an Enterprise feature that runs every 5 minutes in the background. When a problem is detected (service down, high CPU, disk full), Claude analyzes the situation and executes only safe, whitelisted actions: restart a service, clear cache, or send an alert. Every action is logged. Destructive or risky actions are never executed automatically.
What about data privacy?
All data is encrypted in transit and at rest. We never store sensitive data like passwords or API keys. You can delete your data at any time.
Get Started Today
Basic is free. Pro $49/mo. Enterprise $149/mo.
Start with a single VPS at no cost. Upgrade to Pro for full database autopilot, or Enterprise for AI Chat, AI Healer, and Cross-Fleet IP Reputation — all with your own Claude API key.