Link safety in OpenGlob Rooms

OpenGlob Rooms is designed to be a secure environment for real-time collaboration. Because Rooms are often used for quick, temporary discussions with various participants, ensuring that the links shared within these spaces are safe is a core part of our security infrastructure.

Currently, OpenGlob Rooms employs a cautious approach to external navigation. This guide explains how our link protection works today and where it is heading.

Current approach: Manual Verification

At this stage of the OpenGlob Rooms beta, our system is configured to alert users on every external link.

When you click a link that leads outside of the openglob.com domain, a security modal appears. This "speed bump" is designed to prevent accidental navigation to potentially malicious sites and to give you a moment to verify the destination.

Why you see the warning

• Context Awareness: In a fast-moving chat, it is easy to click a link without looking at the URL.
• Identity Protection: Malicious sites often attempt to "spoof" login pages to steal OpenGlob SSO credentials.
• Zero-Footprint Safety: Since Rooms does not store chat history permanently, preventing a breach in real-time is the most effective way to protect your data.

How to identify suspicious links

While the OpenGlob warning acts as a first line of defense, we recommend following these best practices before clicking "Open Link":

1. Check the Domain: Look closely at the URL displayed in the security modal. Malicious links often use "typosquatting" (e.g., openg1ob.com instead of openglob.com).
2. Verify the Sender: If you are in a Room with unverified guests, be more cautious with links than you would be in a private session with known collaborators.
3. Watch for URL Shorteners: Be extra careful with shortened links (like bit.ly or tinyurl), as they hide the final destination until you click them.

Future Roadmap: Smart AI Detection

As the OpenGlob ecosystem grows, we are moving away from universal alerts toward a more seamless, intelligent experience.

AI-Powered Link Analysis

We are currently developing a Smart AI Security Model specifically for Rooms. Once deployed, this system will analyze shared links in real-time to detect:
* Spam Patterns: Automatically flagging and hiding repetitive or "bot-like" link sharing.
* Phishing Signatures: Identifying sites known for credential theft or malware distribution.
* Reputation Scoring: Using global threat intelligence to allow safe, trusted domains to open instantly while blocking high-risk addresses.

What this means for you

In the future, you will only see a warning if our AI model detects something suspicious. Safe, reputable links (like those to major productivity tools or news sites) will open directly, reducing friction without compromising your security.

Reporting Malicious Content

If you encounter a link that you believe is malicious or being used for spam, we recommend that the Host Ban the Participant immediately. This not only removes the user but ensures they cannot rejoin the session to post further links.

For persistent issues, you can submit a report through our feedback system to help train our security models.

Need more help?
Learn about Safety, privacy, and security in Rooms.