Constrain, log and scan your MCP connections for security vulnerabilities.

LLM proxy to observe and debug what your AI agents are doing.

Invariant Documentation

Let Claude control a web browser on your machine.

Helps you build better AI agents through debuggable unit testing

Guardrails for secure and robust agent development

A better way of testing, inspecting, and analyzing AI Agent traces.

Example implementation of MCP Streamable HTTP client/server in Python and TypeScript.

Invariant SDK client library.