Skip to content
SPECIFICTION
Get listed

garak

garak.ai

Free plan WindowsmacOSLinuxSelf-hosted
Visit garak.ai
Spec score 7.9 #9 of 27 in AI Red Teaming Tools

What it is

garak is a command-line scanner for probing language models and systems that use them for security weaknesses. Its checks cover issues such as hallucination, data leakage, prompt injection, misinformation, toxicity generation, and jailbreaks. Probes try to elicit failures, while detectors inspect responses for patterns such as unsafe content or a failure to refuse. A standard scan can run without supervision and produce a report, with scan settings available for customization. Integrations include OpenAI, Hugging Face, Cohere, Replicate, AWS Bedrock, LiteLLM, REST-accessible models, and custom Python integrations. Hugging Face models can run locally or be accessed through an inference API or private endpoint. Output includes screen results, a detailed report log, a hit log, and a debug log. garak assesses text output only. It is developed in Linux and OSX and can be installed with pip or from source; listed platforms include Linux, macOS, Windows, and self-hosted use. It is free, with a free plan and no free trial. Listed capabilities include custom metrics, LLM-as-a-judge, and safety evaluations.

Spec sheet

Purpose
garak is an LLM vulnerability scanner that probes language models and systems using them for security weaknesses.Read 7 Oct 2026 · garak.ai
Deployment
garak is a command-line tool developed in Linux and OSX and can be installed with pip or from source.Read 7 Oct 2026 · github.com
Integrations
The project lists support for OpenAI, Hugging Face, Cohere, Replicate, AWS Bedrock, LiteLLM, REST-accessible models, and custom Python integrations.Read 7 Oct 2026 · github.com
Support
The project points users to its user guide, reference documentation, Discord, GitHub issues, and an NVIDIA email address for support.Read 7 Oct 2026 · garak.ai
Automated scans
garak can run a standard scan and produce a report without supervision, while allowing scan configuration to be customized.Read 7 Oct 2026 · docs.garak.ai
Local and hosted models
The documentation describes running Hugging Face models locally or accessing them through an inference API or private endpoint.Read 7 Oct 2026 · github.com
Maintainer
The project website says garak is backed by NVIDIA and is updated by NVIDIA and the community.Read 7 Oct 2026 · garak.ai
Output scope
The reference documentation says garak supports mainstream any-to-any large models but assesses text output only.Read 7 Oct 2026 · reference.garak.ai
Probe and detector design
Probes attempt to elicit failures, and detectors examine model outputs for failure modes such as unsafe content or failure to refuse.Read 7 Oct 2026 · reference.garak.ai
Reports
garak produces screen output, a detailed report log, a hit log, and a debug log.Read 7 Oct 2026 · docs.garak.ai
Security reporting
The repository directs security vulnerability reports to NVIDIA’s security submission form or [email protected] and says not to report them through GitHub.Read 7 Oct 2026 · github.com
Threat coverage
The project lists hallucination, data leakage, prompt injection, misinformation, toxicity generation, and jailbreaks among the weaknesses it probes.Read 7 Oct 2026 · github.com

Plans and prices

Free Free Free tool · open-source software Read 7 Oct 2026 · github.com

Every garak plan and price →

Strong points and limits

Strong points

  • Scans for prompt injection, data leakage, and jailbreaks.
  • Standard scans can run without supervision.
  • Supports multiple providers and custom Python integrations.
  • Produces detailed, hit, and debug logs.

Limits

  • Assessments cover text output only.
  • No free trial is listed.
  • Installation is by pip or from source.

Who it suits

garak suits teams and practitioners who want to scan language models for security and safety weaknesses. It is particularly relevant to users who can work with a command-line tool and need text-output assessments.

Verdict

garak offers a free, configurable way to probe supported models and inspect scan output. Its text-only scope and command-line deployment are important constraints to weigh.

Where Specifiction places it

Alternatives to garakAll alternatives →