Lakera

Lakera is a security platform for generative AI applications, autonomous agents and employee AI use. It inspects prompts and responses inline to block prompt injection, jailbreaks and data exfiltration before they reach a model or leave the organization, and it applies comparable controls to the tool calls and file access that agents perform at runtime. Policies are managed centrally and applied across applications without code changes, and the product is model agnostic and multimodal, covering chatbots and voice bots alike. The platform is organized into three areas. Workforce AI Security discovers shadow AI across browsers and SaaS tools, detects sensitive pastes and applies context aware data protection with policy controls by user, application and action. AI Agent Security provides real time threat detection and data leakage protection for applications already in production. AI Red Teaming runs direct and indirect attack simulations, ranks vulnerabilities by risk and supplies collaborative remediation guidance. The vendor describes an API first, cloud native architecture, observation of more than 100 languages, and runtime latency under 50 milliseconds even for very large prompts and context windows. Pricing is not published. The site routes buyers to a demo request or a sales conversation, so plan tiers and amounts are quoted rather than listed. One fact is worth knowing before evaluating it. Lakera has been acquired by Check Point Software Technologies, and the integration is already visible on the site: several routes, including the product updates and company pages, now serve Check Point content instead of Lakera content. Expect the technology to be positioned within Check Point's wider AI security portfolio.