The open-source desktop agent that monitors AI tool usage across your organization
Oximy Sensor runs on employee devices and monitors network connections to detect AI tool usage. It captures what AI applications are being accessed—ChatGPT, Claude, Gemini, and 2,200+ other models—without requiring any code changes or SDK integration.
The sensor intercepts TLS at the system boundary, logging:
- Which AI tools and models are being used
- When requests are made
- Which devices are making them
- Request/response metadata
All events are sent to your Oximy dashboard for centralized visibility and governance.
📅 Book a 30-minute demo to see how organizations are using Oximy to govern AI usage across their workforce.
In the demo, we'll show you:
- Live AI activity detection across your tools
- Custom policy enforcement in action
- Real-world governance workflows
- How to go from install to visibility in 90 seconds
- Zero instrumentation: No code changes, no SDK integration
- Comprehensive coverage: 2,200+ AI models, 100+ apps, 500+ websites
- Privacy-first: Configurable data capture policies
- Open standard: Fully auditable, open-source sensor
- 90 seconds to first detection: From install to visibility
- Shadow AI discovery
- Compliance & audit trails
- AI governance & policy enforcement
- Usage analytics & cost tracking
This repository is fully open source so you can verify exactly what runs on your devices. Key areas to review:
| Directory | Purpose |
|---|---|
OximyMac/ |
macOS sensor application |
OximyWindows/ |
Windows sensor application |
mitmproxy/addons/ |
Network interception addons |
registry/ |
AI apps, websites, and model definitions |
The fastest way to understand if Oximy is right for your organization is a quick demo.
Oximy Sensor is built on the shoulders of giants:
- mitmproxy — The powerful open-source TLS interception library that powers our network monitoring capabilities
- models.dev (GitHub) — Open database of AI models and providers that powers our detection coverage
Made with ❤️ in SF for 🌍