March 2026 — A 3-person startup received an $82,000 Gemini API bill in 48 hours after their key was silently reauthorized by Google. Their normal monthly spend was $180. Read the incident → Your GCP API keys have a kill switch. One that can't be turned against you. CloudSentinel monitors your Google Cloud API usage in real-time and auto-revokes keys the instant they breach your threshold — powered by a Revoke-Only IAM role that is structurally incapable of creating or modifying your infrastructure. Request Early Access — 50% Off See How It Works Request-based. Not billing-based. Revoke-Only IAM Auto-revoke within minutes Built for teams that can't afford downtime THE PROBLEM Google Cloud has no kill switch. Every second a leaked key runs costs you money. Runaway bills before you even wake up A single leaked key can rack up $10,000+ in charges overnight. Google's first alert arrives after the damage is done. Quota tools don't revoke. They throttle. Native GCP rate limits slow down requests — they don't kill the key. An attacker just keeps dripping through. Manual monitoring means humans sleep. By the time your on-call dev sees the alert, opens the console, and finds the key — you've already lost. gcp-billing-monitor — live [02:14:33] api_key: AIzaSy••••••••• [02:14:33] requests_today: 1,240 [02:14:34] requests_today: 4,892 [02:14:35] requests_today: 19,442 ⚠ anomaly detected [02:14:35] estimated_bill: $847.00 [02:14:36] estimated_bill: $3,291.00 [02:14:36] estimated_bill: $9,104.00 [02:14:37] STATUS: NO AUTOMATED RESPONSE [02:14:37] STATUS: MANUAL INTERVENTION REQUIRED HOW IT WORKS Revoke-Only by architecture. Not by promise. CloudSentinel doesn't poll your project every minute. It uses your Service Account to create Alerting Policies directly inside your GCP project. When a threshold is hit, Google pushes a Pub/Sub webhook to CloudSentinel. We receive it and fire the DeleteKey command automatically — within minutes of Google detecting the breach. No polling. No delay. No manual step. 01 Grant IAM Role Assign CloudSentinel's read-and-revoke Custom IAM Role to your GCP project. Takes 60 seconds. → 02 We Set the Sensors CloudSentinel creates Alerting Policies inside your project. Google monitors your usage — we just listen. ↓ 03 Instant Revocation Threshold crossed? Google's monitoring detects it and pushes a Pub/Sub alert to CloudSentinel. We call DeleteKey automatically — no human in the loop, no manual step. Even in a full platform breach, an attacker inherits a role that can only remove access — never create infrastructure. This is not a policy. It is a structural absence of permission. Your key is dead before you even read the alert email. We don't rely on billing data — billing updates are delayed by hours. We act on raw request count, updated in near real-time. WHAT YOU GET Three guarantees. Zero surprises. Zero-Liability Security We only see key IDs — never key values. Our role can read metadata, create monitoring rules, and revoke a key. That's it. No ability to create keys or touch anything else. Request Volume Kill Switch Set a request count limit per API key. When it's crossed, CloudSentinel revokes that exact key automatically. No manual step, no human in the loop. Instant Pub/Sub Alerts Google fires an alert the moment your threshold is crossed. CloudSentinel receives it and acts immediately — within minutes, while you sleep. PRICING Simple pricing. No free tier. A professional tool for teams that need it to work. Early access members get 50% off their first 3 months. 🎉 Early Access — 50% off first 3 months. Limited to first 100 teams. Starter $9/mo $19/mo 3 GCP projects 10 API keys monitored Request volume kill switch Request-based threshold per key Email alerts Request Early Access Most Popular Pro $24/mo $49/mo 10 GCP projects Unlimited API keys Request + alert kill switches Webhook event logs Priority email support Request Early Access Business $49/mo $99/mo Unlimited GCP projects Unlimited API keys Full webhook event history Slack + PagerDuty alerts Dedicated support Request Early Access FAQ The questions your security team will ask. We built CloudSentinel for developers who are skeptical by default. Good. Here are the hard questions answered. Why would I trust a third party with IAM access to my GCP project? That's exactly the right question to ask. Most tools request broad OAuth access or Owner-level permissions. CloudSentinel requests a Custom IAM Role with a permission surface of exactly three things: read your API key list, read your usage metrics, and delete a key. The delete permission only works on existing keys — the role does not contain apikeys.create. This isn't a policy we promise to follow. It is a structural absence at the GCP permission level. Even if our entire platform was compromised, an attacker could only remove keys from your project — never create infrastructure, never access your data, never touch anything else. What happens if CloudSentinel goes down? Does my GCP pro...