Product / UX / Safety & Guardrails

garak

An open LLM vulnerability scanner maintained by NVIDIA. It probes models for hallucination, data leakage, prompt injection, misinformation, toxicity, and jailbreaks across many providers (Hugging Face, OpenAI, Bedrock, and others), the nmap of LLM security testing.

garak LLM vulnerability scanner, Apache-2.0, maintained by NVIDIA; ~8.3K GitHub stars (June 2026). Verified live June 2026.

Openness

5 high confidence

5.0

license: Apache-2.0(OSI)
source: public
self-host: yes
service: none
core-gated: ungated

Fully open source: Apache-2.0 scanner, self-hostable, no proprietary tier.

https://github.com/NVIDIA/garak recorded 2026-06-29
Apache-2.0; ~8.3k stars; probes for injection, leakage, jailbreaks, toxicity across providers

Adoption

3 low confidence

3.0

~8.3K GitHub stars (June 2026); stars-only proxy, capped per methodology.

https://github.com/NVIDIA/garak recorded 2026-06-29
~8.3k stars (June 2026)

Capability

4 medium confidence

4.0

Broadest open library of LLM vulnerability probes; multi-provider coverage.

https://github.com/NVIDIA/garak recorded 2026-06-29
probes: hallucination, leakage, injection, misinformation, toxicity, jailbreaks

Unchanged since 2026-08-01 (last edited, not re-checked)