PATIENCE

--mode=adversarial --verbose

Automated red teaming and compliance validation for AI chatbots — purpose-built for regulated industries.

Why Patience?

Zero cloud dependencies. Built for teams where data residency and auditability aren't optional.

100% On-Device

No cloud. No telemetry. Test data, API keys, and results never leave your hardware — satisfying HIPAA, SOC 2, FedRAMP, and GDPR data residency requirements out of the box.

AI-Powered Red Teaming

Purpose-built adversarial engine attacks your chatbot automatically — prompt injection, jailbreaks, edge cases, stress loads. Know your exposure before your users do.

Audit-Ready Evidence

Every test run produces structured reports in JSON, HTML, Markdown, and CSV. Hand auditors proof of AI validation without manual documentation effort.

Works With Your Stack

Connects to OpenAI, Anthropic, Ollama, and any custom endpoint over HTTP or WebSocket. Switch models without changing your test suite.

Three Testing Modes

From pre-deployment validation to autonomous red teaming — one tool covers the full lifecycle.

Scenario Testing

// Before you ship
  • Script multi-turn conversation flows and run them against any endpoint
  • Validate each response against custom business rules in real time
  • Catch regressions before they reach production
  • Generate pass/fail audit records automatically

Conversation Forensics

// After something goes wrong
  • Import production logs in JSON, CSV, or plain text
  • Surface failure patterns, policy violations, and drift over time
  • Filter by date, user segment, or custom criteria
  • Turn incidents into reproducible test cases

Built for Security

Open source, MIT Licensed, and fully auditable — every line of code is visible and verifiable.

Swift 6+
SwiftUI
macOS 26.0+
Async/Await
NaturalLanguage
Keychain Services

Free & Open Source

A lean, native Swift application with zero third-party cloud dependencies. API keys stay in Keychain. The codebase is public, forkable, and ready for your security team to review.

View on GitHub