The AI Security Vendor Directory.
Categorized by primary positioning, not by marketing claims. Free to browse. Free for vendors to claim.
AI Runtime Security and Prompt Injection Defense
Inline protection for LLM applications. Inspects prompts and outputs in real time. Blocks prompt injection, jailbreak, and prompt leakage.
CalypsoAI
Inference platform for enterprise GenAI with policy, observability, and red teaming.
Lakera
Real-time AI security platform focused on prompt injection and content moderation.
Pillar Security
Discovery, posture, and runtime protection for GenAI applications.
Robust Intelligence
Validation and runtime protection for AI applications. Acquired by Cisco in 2024.
WitnessAI
AI policy enforcement and observability for the enterprise.
AI Security Posture Management (AI-SPM)
Discovery, inventory, and posture grading for AI assets across cloud, SaaS, and codebase. The CSPM equivalent for AI.
AI Red Teaming and Adversarial Testing
Adversarial testing platforms for models, applications, and prompts. Continuous or scheduled red teaming. Ships findings with remediation guidance.
AdversaAI
Adversarial ML research and AI red teaming firm.
HiddenLayer
ML model scanning and adversarial detection under the AISec Platform.
Mindgard
Continuous automated AI red teaming spun out of Lancaster University.
SplxAI
Automated red teaming for LLM applications and AI agents.
AI Governance and Risk Platforms
Inventory, policy, framework mapping, and lifecycle governance. Supports NIST AI RMF, ISO 42001, EU AI Act compliance workflows.
Credo AI
AI governance platform aligned to NIST AI RMF and EU AI Act.
FairNow
Practical AI governance focused on hiring tech and HR systems.
Holistic AI
AI audit and governance platform focused on fairness assessment.
Monitaur
AI governance for regulated insurance and financial services.
Saidot
EU AI Act compliance and AI transparency platform.
Securiti AI
Data and AI command center covering governance, mapping, and access controls for GenAI.
Trustible
AI governance for evolving compliance requirements (EU AI Act, NIST AI RMF).
LLM Output Filtering and AI DLP
Data-loss prevention purpose-built for GenAI. Detects sensitive data in prompts and outputs. Integrates with Copilot, ChatGPT, and proprietary LLM apps.
Aporia
ML observability turned LLM guardrails platform.
Lasso Security
GenAI security platform with DLP, runtime, and developer-focused controls.
Nightfall AI
Cloud DLP covering GenAI usage, including Copilot and ChatGPT.
Prompt Security
Inline GenAI security platform spanning employee usage and developer tools.
Synthetic Data and Privacy for AI
PII redaction, synthetic data generation, and privacy vaults that let teams use AI without exposing regulated data.
Cloud-Native AI Security
AI security capabilities offered by hyperscalers and platform vendors as part of broader cloud or security suites.
Amazon Web Services
Bedrock Guardrails and AWS-native AI safety controls.
Cisco AI Defense
Cisco's productized AI security suite combining Robust Intelligence and Hypershield.
CrowdStrike
Falcon for AI Workloads and Charlotte AI for security operations.
Datadog
LLM Observability inside the broader Datadog platform.
Google Cloud
Sensitive Data Protection, Model Armor, and Vertex AI guardrails.
IBM
Guardium AI Security and watsonx.governance for regulated industries.
Microsoft
Purview AI Hub and Defender for Cloud cover Copilot security and AI workloads.
Palo Alto Networks
Prisma AIRS: runtime AI security and Protect AI capabilities consolidated.
LLM Evaluation, Observability and Quality
Test, monitor, and grade LLM outputs in development and production. Hallucination detection, regression testing, traceability, and continuous quality measurement.
Arize AI
ML and LLM observability with the open-source Phoenix framework.
Comet (Opik)
ML experiment platform Comet with the Opik open-source LLM eval tool.
Confident AI (DeepEval)
DeepEval open-source eval framework plus a hosted regression-testing platform.
Galileo
GenAI evaluation, observability, and protection for enterprises.
Helicone
Open-source LLM observability with one-line proxy integration.
HoneyHive
Continuous evaluation and observability for AI products.
Langfuse
Open-source LLM engineering platform. Observability, evals, and prompt management.
LangSmith
LangChain's hosted observability and evaluation platform for LLM apps.
Openlayer
Continuous evaluation and monitoring for AI systems and LLM applications.
Patronus AI
Automated AI evaluation with research-grade benchmarks.
Traceloop
OpenLLMetry-based observability for LLM apps.
WhyLabs
ML and LLM observability built on the open-source whylogs and langkit libraries.
AI Agent and Non-Human Identity Security
Identity, authorization, and lifecycle controls for AI agents and non-human identities (NHIs), including as agentic AI moves into production.
Astrix Security
Non-human identity threat detection and posture for SaaS, cloud, and AI agents.
Britive
Just-in-time access for cloud, SaaS, and AI workloads.
Entro Security
Secrets security and non-human identity for AI and engineering.
Oasis Security
Non-human identity (NHI) management, extending to the AI agent identity layer.
Token Security
Machine-first identity security for non-human identities and AI agents.
Deepfake and Synthetic Media Detection
Detection, attribution, and provenance for AI-generated content. Voice, image, video, and document deepfakes. Used in fraud, KYC, and content authenticity workflows.
Hive AI
AI moderation, content classification, and AI-generated content detection.
Reality Defender
Multi-modal deepfake detection for voice, image, video, and document.
Sensity AI
Deepfake detection and identity-fraud prevention.
Truepic
Content provenance and authenticity at the point of capture.
Open-Source and Frameworks
Open-source projects and reference frameworks used by enterprises and consultancies to harden AI deployments.
Garak
Open-source LLM vulnerability scanner.
Giskard
Open-source LLM testing framework with hosted hub.
LLM Guard
Open-source security toolkit for LLM-powered applications.
NVIDIA NeMo Guardrails
Open-source toolkit for adding programmable guardrails to LLM apps.
Promptfoo
Open-source LLM evaluation, red teaming, and security testing.
PyRIT
Microsoft's open-source Python Risk Identification Toolkit for GenAI.
Claim your listing.
If you represent a vendor in this directory, you can claim the listing: submit corrections, update your product line, or add capability detail. Listings are free. Claimed listings get a verified badge.
Claim a listingSubmit a vendor.
If a notable AI security vendor or product is missing, send it in. Editorial review for fit. Independent: CWS is a services firm, not a vendor.
Submit a vendorEditorial note. Listings are categorized for editorial clarity. Product names belong to their respective owners. Categorization reflects primary positioning at time of writing. CWS is independent and does not accept payment for inclusion or ranking. Categorization is best-effort and based on primary positioning at time of writing.
Choosing among 60 vendors is hard.
The AI Posture Check tells you which dimensions you actually need to address so you can shortlist vendors that fit your gaps, not the buzziest demo.