Hound Shield sits inline between your team and every AI tool — scanning every prompt for CUI, PHI, PII, and secrets in under 10ms, before it reaches ChatGPT, Copilot, or Claude. SOC 2, HIPAA, and CMMC Level 2 enforced simultaneously. One proxy URL. No code changes.
Setup
Point your AI tools at a proxy URL. That's the whole installation. No firewall rules, no network changes, no IT ticket required.
Replace api.openai.com with proxy.houndshield.com in your environment config. Works with ChatGPT, Copilot, Claude, Gemini — any OpenAI-compatible tool. No code changes, no agents to install.
OPENAI_BASE_URL=https://proxy.houndshield.com16 detection engines check each request against CMMC, HIPAA, and SOC 2 patterns in under 10ms. Your team doesn't notice anything different. The models don't know.
Every scan is recorded with a timestamp, matched pattern, and the framework it triggered. Tamper-proof SHA-256 signed logs. Exportable for your C3PAO or auditor on demand.
Compliance
CMMC Level 2
We map all 110 NIST SP 800-171 Rev 2 controls to your actual AI usage. Every blocked prompt records which control it affects. Your SPRS score updates in real time.
HIPAA
Names, SSNs, MRNs, dates of birth, device IDs, IP addresses, and 12 more. If a prompt contains PHI, Hound Shield quarantines it and logs the event. You set the policy; Hound Shield enforces it.
SOC 2 Type II
Every AI prompt stored in an append-only log with a SHA-256 integrity hash. Your auditors get read-only access. No manual exports, no spreadsheets, no guessing what your team sent.
Every LLM request inspected before it leaves the network
CUI, PII, IP, PHI, secrets, CAGE codes, clearances — all matched
SHA-256 tamper-evident logs ready for any compliance audit
Works with ChatGPT, Copilot, Claude, Gemini — all simultaneously
Defense contractors who start now have 6–8 months to close gaps before assessments. The average SPRS score for an unprepared contractor is −68. Hound Shield shows you exactly where you stand — for free.
Get Started Free