Invariant Documentation
Ban words
Initializing search
Invariant Documentation
Getting Started
Guardrails
Guardrails
Introduction
Agents and Traces
Agent Guardrails
Agent Guardrails
Tool Calls
Loop Detection
Dataflow Rules
Code Validation
Content Guardrails
Content Guardrails
Personally Identifiable Information (PII)
Jailbreaks and Prompt Injections
Images
Moderated and Toxic Content
Regex Filters
Copyrighted Content
Secret Tokens and Credentials
Sentence Similarity
LLM-as-Guardrail
Guardrails in Gateway
Guardrails in Explorer
Rule Writing Reference
MCP-Scan
MCP-Scan
Overview
Scanning
Proxying
Guardrails
Issue Codes
Integrations
Integrations
Explorer
Explorer
Overview
Benchmarks
Search
Self-Host Explorer
Traces and Datasets
Explorer API
Explorer API
Client Setup
Trace Format
Annotations
SDK Installation
Dataset Metadata
Dataset Metadata
GetDatasetMetadata API
UpdateDatasetMetadata API
Uploading Traces
Uploading Traces
File Uploads
Push API
Append Messages API
Gateway
Gateway
Overview
Self-Hosted Gateway
LLM Integrations
LLM Integrations
OpenAI
Anthropic
Gemini
Agent Integrations
Agent Integrations
Microsoft AutoGen
OpenAI Agents SDK
OpenAI Swarm
OpenHands
SWE Agent
Browser Use
Testing
Testing
Overview
Quickstart
Examples
Examples
Computer Use Agent
LangGraph
Function Calling Agents
OpenAI Swarm
Running Tests
Running Tests
pytest Compatibility
Visual Debugging
Writing Tests
Writing Tests
Integrate Your Agent
Matchers
Parameterized Tests
Tests
Traces
Ban words