Open-source CLI for evaluating and red-teaming LLM apps: compare models side by side, assert on outputs in CI, and scan agents and RAG pipelines for jailbreaks and data leaks.
Lakera
What is Lakera?
Runtime guardrails for LLM apps and AI agents: an API that screens prompts, tool calls and outputs for prompt injection, data leakage and policy violations.
- Pricing
- Freemium
- Free plan
- Yes
- Category
- Coding & Development
- Made by
- Lakera (Check Point)
- Updated
Lakera screenshots
Lakera features & details
Lakera builds security for generative AI applications. Its best-known product, Lakera Guard, now sold as Check Point AI Guardrails after Check Point acquired the company, is an API that sits between an application and its model. The app sends each user input, reference document, tool response and model output to the Guard API, which returns whether a threat was flagged so the app can block, rewrite or log the interaction.
The guardrails detect prompt attacks (prompt injection and jailbreaks, including indirect attacks hidden in retrieved documents or tool descriptions), leakage of sensitive data and personally identifiable information, offensive or violent content, links outside an allowed list of domains, and off-policy agent behaviour such as dangerous deviations from an agent's mandate or calls to tools that are not allowed. Detection works across more than 100 languages, and teams can add custom detectors for their own policies.
Security teams manage everything from a web dashboard: projects run in Detect or Enforce mode, policies set which detectors apply and how sensitive they are, and the logs view lists every request with the threats found, the policy applied and the latency. A playground with a demo chatbot shows attacks being flagged and blocked before any integration work. The wider AI Agent Security tier adds agent discovery across platforms such as Amazon Bedrock, Microsoft Copilot Studio, Salesforce Agentforce and n8n, with a risk rating for each agent mapped to OWASP and MITRE ATLAS. Lakera also offers AI red teaming.
Accounts are self-serve: anyone can sign up, try the playground and create API keys. Production and enterprise pricing is arranged with sales rather than published.
No reviews yet — sign in to write the first one
Apps similar to Lakera
The closest matches are Promptfoo and ClawSecure. See every Lakera alternative compared by price and free plan.
Security scanner and runtime monitor for AI agents that audits skills, MCP servers and CLI tools for malicious code and prompt injection.
AI-native application security platform that finds code, logic and auth vulnerabilities, filters false positives and writes fixes you merge as pull requests.
All-in-one application security platform that scans code, dependencies, cloud and containers, and turns findings into merge-ready AI AutoFix pull requests.
Developer security platform that scans code, dependencies, containers and IaC, with Snyk Agent Fix generating AI-written fixes for the vulnerabilities it finds.