AI firewalls and guardrails
Inline filters that inspect prompts, responses and tool calls of AI applications an organization builds.
12 vendors · alphabetical, never ranked
- AurascapeAurascape sells software that discovers AI applications an organization uses and builds, and applies real-time data categorization, threat inspection and intent-based controls to those interactions.Privately held
- Enkrypt AIEnkrypt AI sells software that red-teams AI systems for security, safety and brand risks, applies real-time guardrails against threats, and monitors AI risk and compliance.Privately held
- HiddenLayerHiddenLayer sells software that secures machine learning models and AI applications, covering model scanning, red teaming, runtime guardrails and protection for autonomous agents and MCP connections.Privately held
- Impart SecurityImpart Security sells a runtime enforcement platform that blocks attacks against LLMs, AI agents, MCP servers, APIs and web applications, including inline agent behavior controls and MCP policy enforcement.Privately held · primary category
- LakeraLakera sells software that secures generative AI systems: runtime protection against prompt attacks and data leakage for AI agents and applications, a red teaming service that tests AI systems for vulnerabilities, and discovery of employees' use of outside AI tools.Owned by Check Point Software Technologies since 2025 · primary category
- Lasso SecurityLasso Security sells a platform that gives enterprises visibility, control and real-time defense across the AI models they use, the agents they build and the applications they ship, combining automated red teaming with runtime enforcement of security policies.Privately held · primary category
- MindgardMindgard sells software that maps AI systems for exposure, red-teams AI models and agents for vulnerabilities, and applies runtime protection against attacks.Privately held
- Noma SecurityNoma Security sells software that discovers AI agents, models and tools across an organization's environments, manages AI security posture and policy enforcement, and combines red teaming with runtime protection against attacks such as prompt injection and data exfiltration.Privately held
- Operant AIOperant AI sells a runtime security platform that reads the intent behind AI agent actions and blocks, allows or redacts them inline, with dedicated protection for AI agents, MCP servers and APIs.Privately held · primary category
- PangeaPangea sells software that gives security teams visibility into how AI is used across an organization and applies guardrails against threats such as prompt injection and sensitive data leakage in AI applications.Owned by CrowdStrike since 2025
- Prompt SecurityPrompt Security sells software that discovers and governs how employees, developers and AI agents use generative AI, and tests and protects AI applications an organization builds against prompt injection and other runtime threats.Owned by SentinelOne since 2025 · primary category
- WitnessAIWitnessAI sells a platform that discovers AI applications, agents and MCP servers in use across an organization, governs what employees and AI agents can do, and applies runtime security controls against threats such as prompt injection and jailbreaks.Privately held
No vendor pays to be listed or to change what we write. Descriptions are ours, drawn from each company's own site and filings. Corrections: /corrections.