Lakera Guard

📖 What is Lakera Guard?

Lakera Guard is a real-time AI security engine that protects LLM applications from prompt injection, data exfiltration, toxicity, and system prompt leakage. It acts as an inline security API gateway with sub-50ms latency.

💡 Why It Matters

Lakera built Gandalf (the viral prompt injection security game), harvesting millions of real-world attack vectors to build the industry's most robust real-time threat intelligence database for LLM safety.

🎯 Workplace Use Cases

• Sanitizing user inputs to prevent indirect prompt injection attacks in enterprise chatbots
• Securing RAG applications from malicious document injection vulnerabilities
• Monitoring enterprise LLM usage for regulatory compliance and safety violations
• Blocking system prompt disclosure attempts by malicious external users

🚀 How to Use It

Integrate Lakera Guard API into your application code (`lakera.guard(prompt)`), pass user prompts through Lakera before sending to your LLM, and block flagged requests.

✨ Key Features

• Sub-50ms ultra-low latency real-time prompt injection detection API
• Trained on real-world threat dataset from millions of Gandalf game attack prompts
• Data Loss Prevention (PII, secret keys, confidential text)
• Zero data retention mode for strict enterprise compliance

💳 Pricing & Plans

Free Tier: Free Developer plan includes up to 10,000 security scans per month.

Plan	Price	Notes
Pro	$250/month	Includes 100,000 security scans/month, custom threat category rules.
Enterprise	Custom pricing	Volume discounts, VPC/on-prem deployment, custom SLAs, dedicated support.