Our Products
Two open-source tools that work together: Shield defends your AI, Dungeon tests it.
The AI Firewall
Oubliette Shield
Detect, deceive, defend. A 5-stage runtime detection pipeline that blocks prompt injection and jailbreak attacks in under 2ms, then deceives attackers with honeypots and honey tokens.
12 LLM Providers9 SDK IntegrationsMCP ServerF1: 0.98Apache 2.0
$ pip install oubliette-shield The AI Red Team Engine
Oubliette Dungeon
Test, measure, report. Automated adversarial testing with 57 YAML-defined attack scenarios across 8 categories, multi-provider comparison, React dashboard, and CLI.
57 Attack ScenariosMulti-ProviderMCP ServerReact DashboardApache 2.0
$ pip install oubliette-dungeon Better Together
Use Dungeon to attack your AI. Use Shield to defend it. Measure your detection rate, close the gaps, repeat.
$ oubliette-dungeon run --target http://localhost:5000/api/chat