What is Invisible Unicode Detection?

How to protect your Gen AI application against Invisible Unicodes

Contact Support

Case Studies

Blog

Educational Videos

Layerup Security

Knowledgebase

Dashboard

Email Support

Layerup Security utilizes a custom algorithm to detect invisible Unicode characters that could be used to manipulate LLM prompts without user awareness.

Invisible Unicode

Layerup empowers you to develop secure, data-driven Generative AI applications with robust **Observability**, **Evaluations**, and **Security**. Instead of using LLMs on vibes, you can use Layerup to make data-driven decisions while securing against risks such as prompt injection or hallucinations. You can also measure when your model hallucinates and put guardrails in place to ensure reliance.

Introduction

Start securing your AI applications in under 5 minutes!

Quickstart

Welcome to the Layerup Security knowledge base. This section provides an overview of our security features, including guardrails for jailbreaking, prompt injection, abuse, and sensitive data protection.

When PII or sensitive data is received by Layerup Security masking, we'll ensure that none of it is ever sent to a third-party LLM, without any interruption in service.

PII & Sensitive Data Masking

Proactively protect your LLM from prompt injection by escaping all prompts that contain untrusted user input.

Prompt Escaping

Layerup Security uses a custom-trained hallucination model based on curated proprietary datasets to detect hallucinations.

Hallucination

Layerup Security employs a custom model to detect and prevent prompt injection in LLM responses, ensuring the integrity of user interactions.

Prompt Injection

Layerup Security employs a custom model to detect and prevent jailbreaking attempts in LLM responses, ensuring the responsible use of language models.

Jailbreaking

Layerup Security employs a custom model to detect and prevent the exposure of sensitive data in LLM responses, safeguarding user and company information.

Sensitive Data

Layerup Security employs a custom algorithm to detect and prevent abuse at multiple levels using this guardrail.

Model Abuse

Layerup Security employs a custom model to detect and moderate harmful content in LLM responses, ensuring safe and respectful interactions.

Content Moderation

Layerup Security employs a custom-trained model to detect phishing attempts within LLM responses, ensuring the safety and integrity of interactions.

Phishing Detection

Layerup Security employs an advanced model to detect the presence of code within LLM responses, excluding markdown or data formats like JSON.

Code Detection

Layerup Security integrates advanced logging and monitoring mechanisms to oversee the operations of LLMs, ensuring transparency and accountability in user interactions.

Logging and Monitoring

The Layerup Security SDK can be used to intercept threats in both prompts and responses, mask PII & sensitive data, and handle incident management for LLM calls.

Execute pre-defined guardrails that allow you to send canned responses when prompts or responses meet a certain predicate.

Execute Guardrails

Mask sensitive information in your prompts before sending them to a third-party LLM.

Mask Prompt

Escape Prompt

Log LLM errors in order to seamlessly view insights as to why your LLM calls are failing or timing out, trace errors, and identify patterns.

Knowledgebase

​What is Invisible Unicode Detection?

​How to protect your Gen AI application against Invisible Unicodes

What is Invisible Unicode Detection?

How to protect your Gen AI application against Invisible Unicodes