AI Asset Inventory
A single inventory of the LLM endpoints, agents, RAG systems, and other AI assets running in production, auto-discovered where the source system exposes them and customer-declared where it does not.
Last reviewed July 2026
The gap AI asset inventory closes
AI assets accumulate faster than the CMDB tracks them. Shadow LLM apps, undocumented MCP servers, and orphan RAG pipelines turn into the first vulnerabilities an auditor finds.
How Penaxtra delivers AI asset inventory
Penaxtra catalogues 11 AI asset kinds today in a single workspace. Each row links the asset to its scans, findings, framework mappings, and risk score. Auto-pull integrations cover managed and self-hosted vector stores, fine-tune providers, and self-hosted model registries. Managed foundation-model and ML platform services across the major cloud providers auto-discover via read-only role.
AI asset inventory capabilities
Auto-pull integrations for vector databases and fine-tune providers
Cloud AI service auto-discovery via read-only role
AI Application grouping: bundle endpoints + agents + RAG + tools as a single risk-owned unit
Hygiene score: verified + reported within seven days share
CSV and JSON export of the full inventory
API endpoints for asset CRUD + tagging
AI asset inventory compliance mapping
Asset inventory satisfies NIST AI 600-1 MAP-1, ISO/IEC 42001 Annex A.6 (AI asset management), EU AI Act Article 17 (quality management system), and supports the AI-BOM expectation in OWASP Agentic Top 10.