Skip to main content
Know Your AI is a security, evaluation, and monitoring platform built to help teams understand, test, and protect AI systems end-to-end. It combines attack-driven datasets, automated LLM judging, compliance analysis, SDK-based monitoring, and a firewall — all in one place — so teams can ship with confidence. This site contains the product overview, quickstart guides, SDK & CLI documentation, and feature docs to get your team up and running.

Platform quickstart

Create your first workspace, connect a product, and run a security evaluation.

SDK documentation

Install the SDK to monitor, trace, evaluate, and protect your AI apps.

CLI documentation

Run evaluations and inspect results from your terminal.

​
What is Know Your AI?

Know Your AI is an evaluation, monitoring, and compliance platform for AI applications. It centralizes:
  • Workspaces for organizing teams, products, and access controls
  • Products representing the AI systems you evaluate (Website, API, or Streaming API)
  • Datasets from the Marketplace and your own uploads — covering attack prompts, safety tests, and benchmarks
  • Evaluations with configurable LLM-as-Judge scoring and automated compliance analysis
  • Monitoring with SDK-based dashboards for requests, tokens, latency, cost, and errors
  • Tracing with span-tree visualization for debugging AI interactions
  • Firewall for real-time input/output validation with risk categorization
  • Compliance dashboards for CCPA/CPRA violation tracking with evidence trails

​
Who is it for?

  • Product teams validating new model releases and prompt changes
  • ML/AI engineers red-teaming models with attack datasets
  • Security teams testing for jailbreaks, prompt injection, and data extraction
  • Risk & compliance teams tracking CCPA/CPRA, EU AI Act, and NIST AI RMF policies
  • Engineering teams monitoring production AI with SDK integration

​
Core concepts at a glance

Product

Explore the platform modules and how they work together.

Workspace

Understand project boundaries, roles, and access controls.

Evaluation

Learn how LLM-as-Judge scoring and security testing work.

Monitoring

Track live performance with SDK dashboards and tracing.

​
Typical workflow

1

Create a product

Add your AI system as a Website, API, or Streaming API product in your workspace.
2

Select datasets

Choose attack and benchmark datasets from the Marketplace or upload your own.
3

Connect and evaluate

Connect your product endpoint and run security evaluations with LLM-as-Judge scoring.
4

Review results

Analyze security scores, per-prompt pass/fail, compliance reports, and screenshots.
5

Monitor in production

Integrate the SDK to track requests, tokens, latency, and errors in real time.

​
Where to go next

Get started

Launch your first evaluation.

SDK

Monitor and protect AI with the SDK.

CLI

Run evaluations from the terminal.

Datasets

Browse curated attack datasets and methods.

Monitoring

Track model health with SDK monitoring.

Firewall

Block dangerous inputs in real time.