We've been building with AI coding tools and kept running into the same problem: no visibility into what's actually happening.
• Who on the team is running which MCP servers?
• What tools are approved vs. just installed?
• What data is going where?
So we built CyberCage.
MCP Management
• Auto-discovery of MCP servers
• Approval workflows for teams
• Org-wide control down to individual tools
• Integrates with GitHub's MCP Catalog
Operations
• Manage allowed AI apps across your org
• Full audit logs (Splunk integration available)
• Notifications via Slack, Teams, Webex, webhooks
Supported platforms:
• Claude Desktop, Claude Code, Codex CLI
• VS Code, VS Codium, Cursor, Windsurf, Zed, Kiro, Antigravity
• IntelliJ IDEA (beta)
• n8n (native integration)
• macOS, Linux, Windows (beta)
In private beta: On-device network agent for content inspection (PII detection, packet metadata analysis)
Coming soon: BYOLLM for inspection, browser extensions
We're also planning to open-source CyberSmol v1.0, a small model fine-tuned for AI threat detection. Will share more when it's ready.
See it in action: https://youtu.be/Zy7XhkQkUlk
Product Updates (0)
No updates yet. Check back later for updates from the team.