Test your prompt against real AI guardrails.
Paste a prompt — or a model's response — and see live what Anoman would block, mask, or flag. The same engine that guards hundreds of models in the gateway.
In the last 30 days, Anoman blocked 892 attacks and masked 11,762 pieces of PII across all traffic — automatically, on every call.
Nothing you paste is stored. Guardrails run live; no LLM is called.
What just ran
These are the real guardrails, not a mock
Injection detection (an ML classifier plus signatures) and PII masking (including Indonesian NIK and Singapore NRIC) run in the same worker pool that guards live gateway traffic. Content moderation covers English and Bahasa Indonesia. Output DLP scans a model's response and redacts leaked secrets and PII. This demo runs with injection blocking on, at a fixed setting. Accounts score and record prompt injection by default and block it only with injection blocking turned on, so your results can differ. No LLM is called and nothing you paste is stored.
Point one base URL at Anoman and every request is guarded.
Injection, PII, and output-DLP checks on your own traffic in minutes, across hundreds of models, on a Jakarta-hosted gateway. Your settings decide what blocks.