Playground
Try LLMShield live — run guided scenarios or send your own prompts
Guided Scenarios
PII Detection
Send a prompt containing a credit card number and watch the guardrail block it.
Expected: Blocked by guardrail (403)
Smart Caching
Send the same prompt twice — the second response comes from cache with near-zero latency.
Expected: 1st: LLM response — 2nd: cache hit
Model Routing
A short prompt routes to the small model; a long prompt routes to the large model.
Expected: Short → small model, Long → large model
Try It Yourself
Free-form Chat
Send any prompt through the LLMShield gateway
Send a message to get started
0/300