Skip to main content
Category

AI Security Tools

Tools that defend AI systems from attack: prompt injection and jailbreaks, sensitive-data leakage, model theft, and abuse of AI agents. This category covers offensive testing and runtime protection, not policy and compliance.

13 tools·2 with free plan·3 with free trial
Read the editor's roundup →

Looking for model risk, monitoring, and regulatory readiness instead? Browse AI Governance & Compliance

The tools here span the AI attack surface. Red-teaming and adversarial-testing platforms probe models for weaknesses before attackers do; runtime guardrails and prompt firewalls block injection at inference time; model detection-and-response watches deployed ML; and data-loss prevention stops secrets and PII leaking into or out of LLMs. If you need model risk management, bias monitoring, auditing, or regulatory readiness instead, see AI Governance & Compliance.

13 tools reviewed2 with a free plan3 with a free trial12 with an API3 self-hostable4 paid-only

Top picks

01

Security platform for AI models and the ML lifecycle

HiddenLayer is a credible, well-funded specialist in model-centric AI security, and its breadth, discovery, supply-chain scanning, runtime defense, and attack simulation, plus mappings to MITRE ATLAS and OWASP, fit enterprises with mature AI governance needs. Notably, it remains independent as of August 2026 while several competitors (Lakera, Prompt Security, Protect AI, Robust Intelligence) have been acquired, which some buyers will see as a differentiator and others as acquisition risk. Pricing is enterprise custom with no public tiers or free plan, so it is aimed at organizations that can engage in a sales process; confirm scope and pricing directly.
paid·Custom (enterprise)
View tool →
02

End-to-end security for generative AI and agents

Lasso is a credible, GenAI-native security vendor, but it publishes no pricing and sells only through sales-gated enterprise deals, so it is not a self-serve tool. It fits organizations with real LLM and agent deployments that need discovery plus runtime guardrails; smaller teams or those just experimenting with AI will likely find it heavier and more expensive than needed. Evaluate it with a scoped proof-of-value against your own AI inventory.
paid·Custom quote
View tool →
03

AI-native data loss prevention across SaaS, AI apps and endpoints.

Nightfall AI is an enterprise data-security product bought by security and compliance teams, not an individual tool, and adoption typically runs through procurement and security review. Its ML-based detection can reduce false positives versus pure regex, and its coverage of generative-AI data leakage is timely, but pricing is quote-based and effectiveness depends on how well it maps to your specific SaaS stack and workflows. Pilot it against real data flows before committing.
contact·Custom
View tool →

All AI Security

Delve logo

AI-driven compliance automation for SOC 2, HIPAA, ISO 27001 and more.

#compliance automation#soc 2#hipaa#iso 27001
View tool
Nullify logo

An autonomous AI workforce for product security.

#appsec#vulnerability-management#ai-agents#devsecops 1 views
View tool
Knostic logo

Need-to-know access controls for enterprise AI.

#ai-governance#data-security#llm-security#access-control 1 views
View tool
Pillar Security logo

Securing the agentic workforce across the AI lifecycle.

#ai-security#ai-red-teaming#agent-security#guardrails 1 views
View tool
Zenity logo

Security and governance for the AI agent workforce.

#ai-agent-security#ai-governance#aispm#shadow-ai
View tool
PromptArmor logo

AI risk intelligence for third-party AI features.

#ai-risk-intelligence#third-party-risk#prompt-injection#grc
View tool
Harmonic Security logo

Zero-touch data protection for the GenAI era.

#data-loss-prevention#genai-security#shadow-ai#data-protection
View tool
Lakera logo

Real-time AI security guardrails against prompt injection and LLM threats

#ai-security#prompt-injection#llm-guardrails#genai-security 1 views
View tool
Mindgard logo

Automated AI red teaming and security testing for LLMs and AI agents

#ai security#red teaming#llm security#prompt injection
View tool
Harden logo

On-device security layer that inspects coding agent actions before they execute

#agent security#coding agents#on-device#guardrails
View tool

Other AI tool categories

Frequently asked questions

What's the best ai security tool?
HiddenLayer currently leads in community votes. HiddenLayer is a credible, well-funded specialist in model-centric AI security, and its breadth, discovery, supply-chain scanning, runtime defense, and attack simulation, plus mappings to MITRE ATLAS and OWASP, fit enterprises with mature AI governance needs. Notably, it remains independent as of August 2026 while several competitors (Lakera, Prompt Security, Protect AI, Robust Intelligence) have been acquired, which some buyers will see as a differentiator and others as acquisition risk. Pricing is enterprise custom with no public tiers or free plan, so it is aimed at organizations that can engage in a sales process; confirm scope and pricing directly.
Are there free options?
2 tools in this category offer a free plan. Another 3 have a free trial.
How do you rank these tools?
Tools are ranked by a combination of community upvotes, editorial review, and feature breadth. Our editors review pricing and capabilities quarterly.
Can I suggest a tool we're missing?
Yes — submit it here. Our team reviews submissions weekly.