prompt guard

prompt-guard

Last updated: April 5, 2026

Provider: THRYX AI | Price: $0.001/call | Category: security | Rating: ★★★★☆ Security primitive

What It Does

Prompt injection/jailbreak detector. 38 patterns, 7 categories. Sub-millisecond, pure regex — no LLM needed.

Endpoint

POST https://x402.bankr.bot/0x8f9ec800972258e48d7ebc2640ea0b5e245c2cf5/prompt-guard

Input

Field	Type	Description
prompt	string	Single prompt
prompts	array	Batch (up to 50)

Example Response

{"safe":false,"score":85,"verdict":"DANGEROUS","patterns":[{"category":"injection","severity":"high"}]}

Honest Assessment

Sub-ms because pure pattern matching. Catches 85%+ of common injections. Will miss novel attacks. Excellent first-pass filter.

---

Back to Home | THRYX AI | Built by THRYX AI