LLMFence Documentation
Everything you need to integrate enterprise-grade AI safety into your agents in minutes. One API call — comprehensive detection of PII, prompt injection, and hallucination.
Quick Start
Make your first API call in under 5 minutes. No configuration required.
API Reference
Complete endpoint documentation, request/response schemas, and error codes.
Security Guides
Deep dives on PII detection, prompt injection defense, and brand safety.
SDKs & Libraries
Ready-to-use code snippets for Python, Node.js, and raw cURL.
What is LLMFence?
LLMFence is an enterprise AI safety API that acts as a firewall between your AI agent's output and your users or systems. Before any LLM-generated text reaches a real person, external system, or critical database — it passes through our guardrail layer.
A single POST call returns a structured risk assessment in milliseconds, complete with a 0–100 risk score, specific violation flags, and a redacted version of the text safe for logging.
Core Capabilities
PII Detection
Identifies and redacts names, email addresses, phone numbers, SSNs, credit card numbers, and other personally identifiable information.
Prompt Injection Defense
Detects jailbreak attempts, instruction overrides, role-play manipulation, and adversarial payloads targeting your AI systems.
Financial Hallucination Guard
Flags unauthorized financial commitments — promises of refunds, investment returns, pricing that your agent was never instructed to offer.
Brand Safety Enforcement
Catches discriminatory, offensive, biased, or brand-damaging language before it reaches any customer-facing channel.
Custom Policy Engine (Scale)
Define your own compliance rules in natural language. Evaluated against every single request your agents make.
⚡ Ready to start?
Your first 500 API calls are completely free on the Sandbox tier. No credit card required.
Create Free Account