SafePrompt Documentation

LLM security that's zero-config and production-ready

Quick Start

Get up and running in 5 minutes with our simple API integration.

Complete API documentation with request/response examples in multiple languages.

Learn how to implement multi-turn protection, caching, and fail-open strategies.

Detects and blocks attempts to override system instructions

Stops attempts to bypass safety guardrails

Prevents system prompt and sensitive data leakage

Blocks XSS, SQL injection, and template attacks

SafePrompt uses a multi-layer validation pipeline that balances speed, accuracy, and cost through intelligent early-exit optimization.

Instant detection of URLs, IP addresses, and file paths with action-based blocking.

⚡ 0ms🎯 5% of attacks💰 $0 cost

Fast regex detection of known attack patterns (XSS, SQL injection, template exploits).

⚡ <10ms🎯 67% of attacks💰 $0 cost

Fast AI model analyzes prompts that pass pattern detection for semantic threats.

⚡ Fast response🎯 Most AI validations

High-accuracy model handles edge cases and ambiguous prompts when Pass 1 is uncertain.

⚡ Thorough analysis🎯 Complex cases only