Darkhunt AI Security vs garak
Darkhunt AI Security
darkhunt.aigarak
garak.aiSide by side
| Figure | Darkhunt AI Security | garak |
|---|---|---|
| Price from | Free | Free |
| Free plan | Yes | Yes |
| Free trial | — | — |
| Open source | — | — |
| Web | Yes | — |
| Windows | — | Yes |
| macOS | — | Yes |
| Linux | — | Yes |
| API | Yes | — |
| Self-hosted | Yes | Yes |
| Plans published | 3 | 1 |
| Spec score · AI Red Teaming Tools | 7.9 (#8) | 7.9 (#9) |
| Deployment | Darkhunt says it supports cloud, hybrid, and self-hosted deployment. | garak is a command-line tool developed in Linux and OSX and can be installed with pip or from source. |
| Integrations | The site lists OpenAI, Anthropic, Azure, AWS Bedrock, Gemini, and self-hosted systems, and says it offers pre-built integrations for agent frameworks, model providers, SIEM, and cloud platforms. | The project lists support for OpenAI, Hugging Face, Cohere, Replicate, AWS Bedrock, LiteLLM, REST-accessible models, and custom Python integrations. |
| Purpose | Darkhunt continuously tests, monitors, and protects AI agents in production. | garak is an LLM vulnerability scanner that probes language models and systems using them for security weaknesses. |
| Support | The free plan includes community support, pay-as-you-go includes dedicated support, and custom annual includes dedicated support and onboarding. | The project points users to its user guide, reference documentation, Discord, GitHub issues, and an NVIDIA email address for support. |
Plans
Darkhunt AI Security
- FreeFree
- Custom AnnualPrice not published
- Pay as you goPriced on request
In their own words
Darkhunt AI Security
Darkhunt AI Security continuously tests, monitors, and protects AI agents in production. It discovers models, agents, and MCP servers across SaaS, cloud, endpoints, and code, including sanctioned and shadow AI. Its observability tools log and replay activity, making agent prompts, responses, and tool calls searchable. Automated red teaming tests agents with their actual tools, permissions, and data access for prompt injection, jailbreaks, and data exfiltration. The platform also checks decision integrity, including manipulated figures, suppressed alerts, RAG poisoning, and confident hallucinations, and traces answers to source data points. Findings can become inline runtime policies, while an MCP gateway proxies requests and responses. Listed integrations include OpenAI, Anthropic, Azure, AWS Bedrock, Gemini, and self-hosted systems. Deployment options are cloud, hybrid, and self-hosted. Findings or controls can map to several named frameworks, including the EU AI Act and NIST AI RMF. The free plan costs nothing and includes full platform access, a free credit allotment, and community support. Paid options have unlisted prices.
garak
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.