About FilterPrompt
FilterPrompt is built by an LLM security research team focused on prompt injection, PII redaction, and the OWASP LLM Top 10. We ship the proxy, the detectors, and the audit tooling as one product.
What we believe
LLM apps need the same input/output controls every other production system has. The right place for those controls is a proxy in front of the model — not buried inside application code that ships with every release.
What we ship
- Provider-agnostic OpenAI-compatible proxy
- Layered detection pipeline (rules + ML + structural)
- Multi-tenant gateway with per-tenant quotas, rules, and IP policy
- Audit logs and verdict replay
- Open benchmark methodology