quanterios
Get started

AIBOM Discovery

Inventory every model, agent, MCP server, dataset, and prompt template.

What this module does

AIBOM Discovery is the system of record for the AI estate. Continuous discovery across first-party models, third-party APIs, on-device models, agentic frameworks, MCP servers, training and inference datasets, prompt templates, and the relationships between them.

Capabilities

Four operational capabilities.

Capability · 01
Models · agents · MCP

Built and third-party models · LLMs · embeddings · agentic systems · MCP-server topology · invocation history.

Capability · 02
Datasets · prompts

Training and inference datasets · prompt-template registry · version pinning · lineage.

Capability · 03
Relationship graph

Which agents call which MCP servers, with which scope, against which underlying models.

Capability · 04
Continuous discovery

Real-time delta detection as the estate evolves, new models, new agents, new MCP servers all surfaced.

Technical detail

What evaluators want to know.

Connectors
OpenAI · Anthropic · Hugging Face · Bedrock · Vertex AI · MCP servers · agentic-framework hosts.
Asset metadata
Model version, agent role, MCP scope, dataset lineage, prompt-template version, owner team, deployment region.
Query API
GraphQL and REST APIs · webhook events for new models, new agents, new MCP servers.
Region pinning
All AIBOM data region-pinned at write time.