Lakera Guard
Lakera Guard is an API that screens prompts, tool calls and model outputs for prompt attacks, data leakage and unwanted content.
What Lakera Guard does
Lakera Guard is a screening service that sits between an application and a language model. The application sends each user input, tool call, tool response and model output to the Guard API, which reports whether it found a prompt attack, leaked personal or sensitive data, content that breaks a moderation policy, a link outside an allowed list of domains, or an agent action outside the tools it is allowed to call. What is checked is set per project through a policy, and a project mode decides whether a detection is only logged or flagged for the application to act on.
The hosted service has a dashboard with analytics, request logs and policy management. A self-hosted container is available to Enterprise customers.
Lakera has been part of Check Point since October 2025, and the vendor's documentation calls the product Check Point AI Guardrails. It is offered on its own or as the runtime layer of Check Point AI Agent Security, which adds agent discovery and risk assessment.
Lakera Guard: owner, license and hosting
| Fact | Value | Source and date checked |
|---|---|---|
| Owner | Check Point Software Technologies Ltd., which closed its acquisition of Lakera on 22 October 2025. | SOURCE ↗ |
| License | Commercial, not open source. Self-hosting requires "a valid AI Guardrails Enterprise license". | SOURCE ↗ |
| Self-hosting | Enterprise only. Kubernetes (Helm), Docker, or air-gapped; the dashboard is not available when self-hosted. | SOURCE ↗ |
| Hosted regions | Requests are processed in the EU, the USA or Singapore, chosen by API endpoint. Logs are stored in one region, EU by default, fixed when the organization is created. | SOURCE ↗ |
Each value was read from the vendor's own page on the date shown. How listings are checked
Lakera Guard plans and prices
Lakera Guard pricing, checked on . Lowest paid plan: Not public. Enterprise is listed as "Let's chat!", through the sales team. Free tier: Community: 10k requests a month, prompts up to 8k tokens, hosted only. Pricing model: A free Community plan capped by monthly requests, and an Enterprise plan with a negotiated package of requests per month.
| Plan | Price | Included usage | Beyond that | Data retention | Users |
|---|---|---|---|---|---|
| Community | $0 | 10k requests a month | – | – | – |
| Enterprise | Not public: "Let's chat!" | Flexible | – | – | – |
The page says: "The default pricing plan (Community) provides free access to a limited number of requests. Contact sales to update your plan and get access to advanced features." Maximum prompt size is 8k tokens on Community and "Configurable" on Enterprise. Hosting is SaaS on Community and "SaaS or Self-hosted" on Enterprise. Data residency is listed as EU on Community and EU/US on Enterprise. SSO, role-based access control and SIEM integration are Enterprise only. The page does not state retention or a number of users for either plan.
As stated on the vendor's pricing page on . A dash means the page does not say.
Lakera Guard pricing history
Lakera Guard's pricing page is not watched automatically. The plans above carry the date they were last checked by hand.
The pricing page has been checked daily since . No price change has been recorded since then.
Where Lakera Guard fits
Teams that want to add screening for prompt injection and data leakage to an existing application with one API call per interaction, without running detection models themselves. The Community plan is free for 10k requests a month, and the processing region (EU, USA or Singapore) is chosen by which endpoint the application calls.
Limits of Lakera Guard
It is a closed, commercial service: there is no open-source version, the Enterprise price is not published, and self-hosting needs an Enterprise license and gives no dashboard. The Community plan is capped at 10k requests a month and 8k tokens per prompt, with SSO, role-based access and SIEM export reserved for Enterprise. Every screened interaction adds a network call to the request path. Its pricing page is drawn by scripts, so this site cannot watch it for changes.
Alternatives to Lakera Guard
-
Amazon Bedrock Guardrails
Amazon Bedrock Guardrails is a managed AWS service that filters prompts and model responses against content, topic and sensitive-data policies.
-
Arthur Evals Engine
Arthur Evals Engine is an open-source service you host that applies guardrail rules to LLM prompts and responses and collects traces for evaluation.
-
Google Model Armor
Model Armor is a Google Cloud service that screens LLM prompts and responses for prompt injection, harmful content, sensitive data and malicious URLs.
-
NVIDIA NeMo Guardrails
NVIDIA NeMo Guardrails is an open-source Python library for adding programmable input, dialog, retrieval and output checks to LLM applications.
-
OpenAI Guardrails
OpenAI Guardrails is an open-source Python library that wraps the OpenAI client and runs configurable checks on model inputs and outputs.
-
Presidio
Presidio is an open-source Python framework that detects and anonymizes personal data in text, images and structured data.
-
SignalVault
BUILT BY THIS SITE'S EDITOR
SignalVault is a hosted proxy that logs requests to OpenAI and Anthropic models and blocks or redacts sensitive data in them.