Understanding the CalypsoAI Dashboard: A Quick Guide
Welcome to your CalypsoAI dashboard—your mission control for LLM security! Here’s a breakdown of what you’ll find, what it all means, and how it helps you keep your organization’s AI usage safe and under control.
Dashboard Overview
The dashboard gives you a bird’s-eye view of your organization’s interaction with large language models (LLMs). You’ll see real-time stats, trends, and security actions, including:
- Blocked Prompts: The number of prompts CalypsoAI flagged and blocked in the last 30 days. (Because sometimes, “No” is the safest answer.)
- Prompts Sent: Total prompts your team sent to LLMs.
- Prompts Sent and Blocked (%): The percentage of prompts that were intercepted for security reasons. A high number here means your security settings are working overtime.
- Most/Least Used Model: Find out which LLMs your team loves—and which ones are gathering digital dust.
Security Scanners
CalypsoAI uses several “scanners” to keep things safe:
- System Prompt Scanner: The MVP, blocking over 37% of risky prompts.
- Prompt Injection Scanner: Keeps an eye out for sneaky prompt attacks.
- Obfuscation & Credit Card Number Scanners: Because nobody wants a data leak.
User & Provider Insights
- Most Blocked Prompts per User: See who’s tripping the alarms most often (looking at you, Allan Healy).
- Latency per LLM Provider: Track how fast (or slow) each provider responds.
- Usage Trends: Discover peak usage hours—perfect for planning (or avoiding) system maintenance.
- Usage per LLM Provider: See which AI models are getting the most action.
Exporting Data
For the data nerds (no shame!), you can download reports in SVG, PNG, or CSV format for deeper analysis.
TL;DR
The CalypsoAI dashboard is your one-stop shop for monitoring AI usage, security events, and team behavior. It’s like having a security guard, traffic cop, and data analyst—all rolled into one handy interface.
Still have questions? Reach out to your admin or the CalypsoAI support team for more details!