Vettd provides a tool that inventories every AI agent, model, server, prompt, and credential in an organization, creating a signed, version‑controlled AI bill of materials (AI‑BOM). It adds security grades, performance metrics, and compliance evidence mapped to standards like NIST AI RMF and the EU AI Act, all generated locally without data leaving the host.
Funding
Funding not disclosed
Founders
Product
Problem
Enterprises deploying AI agents and autonomous workflows struggle to maintain visibility into every component—agents, models, prompts, servers, and credentials—making it difficult to assess security, performance, cost, and regulatory compliance. Existing tools often rely on external reports or PDFs, providing limited real‑time assurance and exposing organizations to supply‑chain risks.
Solution
Vettd offers a fully local AI asset management platform that scans an organization’s entire agentic estate and generates a signed AI‑Bill of Materials (AI‑BOM). The scanner inventories every agent, model, MCP server, prompt, and credential, assigning versioned identifiers and a security grade derived from existing tooling. Performance and cost grades are calculated from workload benchmarks, usage telemetry, and provider invoices, giving a holistic view of asset value. All evidence is mapped to standards such as NIST AI RMF, ISO 42001, the EU AI Act, and OWASP, producing verifiable compliance artifacts that remain on‑premises until the user chooses to share them. The platform operates with zero egress, ensuring that no data leaves the host during scanning.
Target Audience
Primary customers are security and compliance teams in enterprises that deploy autonomous AI agents, as well as auditors and AI product developers needing a trustworthy, on‑premises inventory and compliance evidence.
Features
- Local, MIT‑licensed scanner (vettd‑cli) that walks the entire AI estate and produces a signed, versioned AI‑BOM
- Asset‑level security grading (credential scope, egress, injection surface, supply‑chain posture) integrated from trusted security tools
- Performance (benchmark against internal workloads) and cost grading using provider invoices and runtime telemetry
- Evidence mapping to NIST AI RMF, ISO 42001, EU AI Act, and OWASP for standards‑based compliance verification
- Zero‑egress operation: all scanning and evidence generation stay on the host unless explicitly exported
- Open, vendor‑neutral schema with signed artifacts that can be audited without relying on PDFs
- Automated remediation actions (revoke credential, reduce scope, quarantine process, roll back version, etc.) based on grade findings