Lakera

What it does

Lakera is an AI-native security platform for enterprise GenAI applications, agents, and MCPs. Its Guard API takes user, LLM, and tool messages and returns a flagged response when threats are detected, with a companion endpoint giving detailed detector results. Prompt attacks, data leakage, and calls to denied tools are screened against a configurable policy.

The platform also includes AI Red Teaming, which runs adversarial tests against AI applications and agents, and Workforce AI Security, which discovers shadow AI and applies policy to employee prompts across browsers, desktop agents, and developer tools. Teams call the Guard API over HTTP from their application and manage policies and projects centrally rather than changing application code.

Lakera suits security and engineering teams protecting LLM-powered applications and organizations governing employee AI use. The Guard SaaS API requires an API key kept out of client-side code, while self-hosted deployments do not use authorization. Pricing is not published, so cost and plan limits must be confirmed with the vendor.

What it can do

• Real-time screening of user and LLM messages for prompt injection, jailbreaks, and data leakage
• Screening of tool responses, tool descriptions, and tool calls, with denial of calls to disallowed tools
• Central policy control by user, app, and action without code changes
• Automated AI red teaming for safety, security, and responsible AI risks
• Shadow AI discovery and context-aware data protection in employee prompts across browsers, desktop apps, and IDEs

Before you choose

• API keys are secrets and must not be shared or placed in client-side code.
• Regional endpoints affect latency; the default api.lakera.ai routes to the nearest region.
• Pricing and plan limits are not published; verify costs with the vendor before committing.

Listed under

• Security
• AI and agent security
• AI security
• Developers
• Enterprise
• Guardrails
• LLM apps
• Prompt injection

Get started

The homepage offers a 'Start for free' call to action and a 'Talk to Sales' option.

1. Create or sign in to a Lakera platform account at platform.lakera.ai.
2. Generate an API key in the API access section of the AI Guardrails Dashboard and send it as a Bearer token in the Authorization header.
3. Send text to the /guard endpoint from your application, choosing a regional endpoint if data location matters.

The site offers a Start for free call to action and a Talk to Sales option. Confirm any free tier, trial terms, and prices with the vendor.

We couldn’t find published prices for Lakera.

Sources for this overview

Product pages, documentation, and pricing checked 30 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Lakera: The AI-Native Security Platform to Accelerate GenAI ↗www.lakera.ai
• API Overview | Check Point AI Security ↗docs.lakera.ai
• Workforce AI Security | Lakera, a Check Point Company ↗www.lakera.ai
• AI Red Teaming: Adversarial Testing for AI Applications and Agents ↗www.lakera.ai
• Lakera: The AI-Native Security Platform to Accelerate GenAI ↗www.lakera.ai

Company website

lakera.ai: product, pricing and policy pages

GitHub

Public code and releases

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

manual_review

Last checked 30 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.