What it does
A single-call guardrail API for detecting prompt injection, jailbreaks, data leaks, and unsafe content with Jev.
Apps & data pipelines
jevegis
Bookmark: jevegis
Keep this in your collection.
Leave a noteWhat would you try with this?
: jevegis
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
How you can use it
A developer can adapt this open code to build a safety filter for an interactive app. The app sends a user message to an outside AI service called TypeSafe. That service checks the text for hidden instructions or leaked personal details. It then returns a score marking if the message seems dangerous.
The setup requires an access key for TypeSafe and a database connection to store scan records. Because the app sends user text to an external service for review, consider what private data you share. The filter blocks messages marked dangerous, but the AI will not catch every trick.