FilterPrompt — AI Security Testing & LLM Vulnerability Scanner
FilterPrompt finds vulnerabilities in your AI applications before attackers do. Connect your LLM app, chatbot or AI agent, run 1,000+ real-world adversarial prompts, and get a detailed vulnerability report with fixes — no source code or infrastructure changes required.
What FilterPrompt tests
- Prompt injection — direct and indirect (RAG, documents, tool results)
- Jailbreak attacks — role-play, DAN-style and obfuscated bypasses
- Sensitive data leakage — PII, secrets and internal data in responses
- System prompt extraction
- Unsafe or harmful AI responses
- OWASP LLM Top 10, MITRE ATLAS, NIST AI RMF, EU AI Act and ISO/IEC 42001 mapping
How it works
Connect any LLM — OpenAI, Anthropic, Google Gemini, Azure OpenAI or a custom endpoint — with your own API key. FilterPrompt sends adversarial probe batteries, grades every response, and produces a report with per-probe evidence, severity and a prioritized fix list (system-prompt patches, refusal templates, guardrail rules). Fix, then rescan to confirm.
Pricing
First scan free, no credit card. Then $36 per full scan — pay per scan, no subscription, credits never expire.