Cloud.ru Open-Sources Guardrails Filter to Protect Data in LLM Requests

In July 2026, Cloud.ru open-sourced Guardrails Filter, a transparent reverse proxy that strips sensitive data from requests to language models and restores it in the responses. The tool is published as open source and is free to use.

What Guardrails Filter does

Guardrails Filter works as a transparent reverse proxy between a client application and an LLM provider. It intercepts a request before it reaches the model, finds sensitive data in it, and replaces that data. In the model's response, it performs the reverse operation and puts the original values back. Neither the application nor the model notices, and the integration code does not need to be rewritten.

The key point is that data never leaves the perimeter in plain form. The model sees anonymized text, while the user gets a normal response with the original values restored.

• Developer — Russian cloud provider Cloud.ru
• Type — transparent reverse proxy
• Function — removes sensitive data from the request and restores it in the response
• Compatibility — works with any LLM provider
• License — open source, free to use

Why businesses need it

Guardrails Filter addresses a problem familiar to any company that connects cloud LLMs to its workflows: prompts inevitably end up containing customers' personal data, contract numbers, and internal documents. All of it goes to the model provider, and the sender can no longer control what happens to that data next.

The proxy breaks that dependence. It adds a layer between the application and the model that guarantees only anonymized text leaves the company. This removes some of the risks tied to personal data protection requirements (152-FZ) and corporate security policies. It also lets teams use even external models in places where the security department previously blocked any data transfer.

Why a cloud provider is giving the code away

Guardrails Filter is published in full, with no cut-down free version and no paid edition "for the real thing." According to one of its developers, Cloud.ru is deliberately sharing the work with the market, and the article promises to explain the motivation behind the move.

"I am one of the developers of this tool and would like to explain in more detail... to answer the question: why would a cloud give away its work to the market for free?" — one of the developers of

Guardrails Filter, in a post on Habr.

The logic for a cloud provider is straightforward: the easier and safer it is for companies to connect LLMs, the more workload they move into the cloud. An open-source tool lowers the barrier to entry and doubles as a showcase of Cloud.ru's engineering expertise.

What this means

Anonymizing data before it is sent to an LLM is moving from homegrown workarounds to a ready-made, reusable component. For teams that security requirements have held back until now, there is a free way to connect cloud models without handing them confidential information. They can also see exactly how the filtering works by reading the source code.

Frequently asked questions

What is Guardrails Filter?

It is a transparent reverse proxy from Cloud.ru that sits between a client and an LLM provider. It removes sensitive data from requests to the model and restores it in the responses, without requiring changes to the application's integration code.

Is Guardrails Filter paid?

No. Cloud.ru released the source code publicly in July 2026. The tool is free and open source, and it works with any LLM provider.