Cybersecurity & SOCLlama Community LicenseWorks with local AI
PurpleLlama
Meta's AI safety models, prompt-injection detection and security benchmarks.
Free facts
- Free tier
- Evals and CodeShield: MIT. Llama Guard, Prompt Guard: Llama Community licenses
- Card required
- No
- Your data
- Runs on your computer
- Setup
- Install & self-host
- The catch
- Benchmarks test an LLM you supply, through a provider API key or a self-hosted model
Last checked Oct 30, 2026
Free facts from docs
- Level
- Basic
- Paid service
- None
- Helps you
- Check AI models and their code for safety and cybersecurity risks
Read from the official docs: github.com/meta-llama/PurpleLlama#readme, github.com/meta-llama/PurpleLlama/blob/main/CybersecurityBenchmarks/README.md
About PurpleLlama
Meta's AI safety models, prompt-injection detection and security benchmarks.
- License
- Llama Community License
- Category
- Cybersecurity & SOC
- Source
- meta-llama/PurpleLlama
Every fact on this page was checked by hand. How we check tools
Alternatives
- AiSOCSelf-hosted AI SOC: alert merging, AI triage and MITRE ATT&CK investigations.
- HexStrike AIConnects AI agents to 150+ existing security tools through MCP.
- SOC RangeSOC analyst training simulator with an optional local AI Triage Copilot: alert summaries, MITRE ATT&CK mapping and next steps, verified against intel.