Cyber Profiler checks prompts for attacks before they reach your model.
Cyber Profiler checks each prompt for prompt injection, jailbreaks and other malicious intent before it reaches your LLM or agent. You send it a prompt, and it tells you whether it's a threat, how likely that is, and how sure it is.
- Quickstart: get a key and analyse your first prompt with the Python SDK.
- Analyse a prompt: the API reference, where you can try it.
- Python SDK reference: every option in the Python SDK.
To get access, visit the Cyber Profiler page.