Block 'Dangerous AI' with Zero On-site Burden. AI Security 'Mamoraku AI' Beta Version Launched.
MONO BRAIN has launched the beta version of 'Mamoraku AI,' an AI security agent that detects shadow AI and automatically masks sensitive information via a browser extension. It offers real-time DLP, shadow AI detection with nudges to safe tools, zero-data exposure through local processing, and an admin dashboard, aiming for a balance between governance and user productivity.
📋 Article Processing Timeline
- 📰 Published: April 15, 2026 at 19:20
- 🔍 Collected: April 15, 2026 at 11:01
- 🤖 AI Analyzed: April 15, 2026 at 17:20 (6h 19m after Collected)
MONO BRAIN Co., Ltd. (Headquarters: Shibuya-ku, Tokyo, Representative Director: Masaki Kato, hereinafter 'MONO BRAIN'), a developer of AI security platforms, has launched the beta version of 'Mamoraku AI,' a next-generation AI security agent that realizes shadow AI usage detection and automatic masking of confidential information simply by adding an extension to employees' browsers.
■ Development Background: Balancing 'Governance (Protection)' and 'On-site Productivity (Ease)'
As the business use of generative AI, including ChatGPT, rapidly progresses, companies are facing new security challenges such as 'inputting confidential information into prompts (data leakage risk)' and 'using unapproved AI tools (shadow AI)'
To solve these issues, MONO BRAIN developed the browser extension-type agent 'Mamoraku AI' with the concept of 'enforcing strong control without compromising on-site usability.'
■ Main Features of 'Mamoraku AI'
No complex system construction required. By distributing the browser extension via MDM, etc., a company-wide AI governance environment can be established in as little as one day.
1. Real-time DLP and 'Automatic Masking (Proposal Type)'
Immediately before an employee sends a prompt to AI, data is scanned in real-time within the browser. If confidential information such as credit card numbers or email addresses is detected, transmission is temporarily paused, and a proposal is made to send it after masking with '*****' with a single click. This allows employees to use AI safely without the hassle of rewriting prompts. It also includes a 'Self-Override Function' that allows users to send the prompt as is after entering a reason, to prevent work stoppage.
2. Shadow AI Detection and 'Guidance to Safe Tools (Nudge)'
Detects and blocks access to unapproved free AI tools, etc., that the company has not authorized. Instead of simply prohibiting, it provides a link to a safe alternative (nudge) with the message 'Please use the in-house standard AI tool instead,' guiding users to the correct tool without causing confusion.
3. Zero-Data Exposure (Completely Local Processing)
All confidential information scanning of prompts is completed within the employee's browser (local environment). The raw input prompt data is never sent or stored on the Mamoraku AI management server, thus not introducing new leakage risks (third-party risks) for implementing companies.
4. Real-time Risk Visualization (Administrator Dashboard)
Information system department administrators can grasp 'DLP activation counts,' 'protected transmissions,' and 'unapproved AI access attempts' in real-time through a dashboard. This visualizes the health of company-wide AI usage while maintaining on-site convenience.
■ Pre-registration for Beta Version
We will begin accepting applications for the closed beta version of 'Mamoraku AI' starting today. Companies struggling with 'wanting to promote on-site AI utilization but being held back by security requirements' or 'high operational load of existing security tools' are encouraged to inquire.
▼ 'Mamoraku AI' - AI Security That Doesn't Stop the Field
Pre-registration for Beta Version / Demo application here:
https://modelsafe.jp/contact
■ About MONO BRAIN Co., Ltd.
MONO BRAIN is a group of AI security specialists aiming to implement 'proactive AI governance that maximizes AI value' in society. Through the development and provision of the integrated platform 'MODEL SAFE,' which protects systems from AI-specific vulnerabilities, and participation in the General Incorporated Association AI Governance Association, we contribute to creating a society where companies can accelerate innovation more boldly.
▼ Overview of AI Security Platform Here:
https://modelsafe.jp/
Keywords: None
■ Development Background: Balancing 'Governance (Protection)' and 'On-site Productivity (Ease)'
As the business use of generative AI, including ChatGPT, rapidly progresses, companies are facing new security challenges such as 'inputting confidential information into prompts (data leakage risk)' and 'using unapproved AI tools (shadow AI)'
To solve these issues, MONO BRAIN developed the browser extension-type agent 'Mamoraku AI' with the concept of 'enforcing strong control without compromising on-site usability.'
■ Main Features of 'Mamoraku AI'
No complex system construction required. By distributing the browser extension via MDM, etc., a company-wide AI governance environment can be established in as little as one day.
1. Real-time DLP and 'Automatic Masking (Proposal Type)'
Immediately before an employee sends a prompt to AI, data is scanned in real-time within the browser. If confidential information such as credit card numbers or email addresses is detected, transmission is temporarily paused, and a proposal is made to send it after masking with '*****' with a single click. This allows employees to use AI safely without the hassle of rewriting prompts. It also includes a 'Self-Override Function' that allows users to send the prompt as is after entering a reason, to prevent work stoppage.
2. Shadow AI Detection and 'Guidance to Safe Tools (Nudge)'
Detects and blocks access to unapproved free AI tools, etc., that the company has not authorized. Instead of simply prohibiting, it provides a link to a safe alternative (nudge) with the message 'Please use the in-house standard AI tool instead,' guiding users to the correct tool without causing confusion.
3. Zero-Data Exposure (Completely Local Processing)
All confidential information scanning of prompts is completed within the employee's browser (local environment). The raw input prompt data is never sent or stored on the Mamoraku AI management server, thus not introducing new leakage risks (third-party risks) for implementing companies.
4. Real-time Risk Visualization (Administrator Dashboard)
Information system department administrators can grasp 'DLP activation counts,' 'protected transmissions,' and 'unapproved AI access attempts' in real-time through a dashboard. This visualizes the health of company-wide AI usage while maintaining on-site convenience.
■ Pre-registration for Beta Version
We will begin accepting applications for the closed beta version of 'Mamoraku AI' starting today. Companies struggling with 'wanting to promote on-site AI utilization but being held back by security requirements' or 'high operational load of existing security tools' are encouraged to inquire.
▼ 'Mamoraku AI' - AI Security That Doesn't Stop the Field
Pre-registration for Beta Version / Demo application here:
https://modelsafe.jp/contact
■ About MONO BRAIN Co., Ltd.
MONO BRAIN is a group of AI security specialists aiming to implement 'proactive AI governance that maximizes AI value' in society. Through the development and provision of the integrated platform 'MODEL SAFE,' which protects systems from AI-specific vulnerabilities, and participation in the General Incorporated Association AI Governance Association, we contribute to creating a society where companies can accelerate innovation more boldly.
▼ Overview of AI Security Platform Here:
https://modelsafe.jp/
Keywords: None