AuditaAI

Navigate the evolving AI security market

Track and compare AI security vendors, products, and documented capabilities. Explore market coverage, find products by security need, and see how offerings relate to widely used security and AI governance frameworks.

Find products by capability

Start with an AI security need and discover products with documented capabilities that address it.

Open

Compare vendors and offerings

Compare products side by side across capabilities, coverage, and supporting product information.

Open

Review market coverage

See how product families cover AI security capability categories across the market.

Open

Explore framework relationships

See how product capabilities relate to OWASP, MITRE ATLAS, NIST AI RMF, ISO/IEC 42001, and NCSC/CISA secure AI development guidance.

Open

Recent additions

Showing last 14 days

Vendor coverage maintenance & updates

Added 7 Sep 2026

Updated OpenAI with Frontier security evaluation following its Promptfoo acquisition, Cyera with Agent Guardian and Ryft agentic data governance, and restored 0DIN to active coverage.

Open research

Four AI security vendors onboarded

Added 7 Sep 2026

Promptfoo, Harmonic Security, Dynamo AI, and Enkrypt AI were added with reviewed evidence for automated red teaming, shadow AI discovery, prompt DLP, runtime gateway proxies, and agent tool governance.

Open research

Teleport agentic AI & MCP governance added

Added 7 Sep 2026

Teleport was added with reviewed evidence for MCP reverse proxying, granular tool-filtering RBAC, and Beams Firecracker microVM agent execution isolation.

Open research

How to evaluate

1. Start with the control need

Use Navigator to find the normalized capability or framework context relevant to your decision.

2. Inspect the offering

Read the behavior claim and its source before treating a capability as relevant to your environment.

3. Apply local validation

Confirm deployment, licensing, integration, and operational effectiveness with current vendor documentation and your own testing.