Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.xpertai.cn/llms.txt

Use this file to discover all available pages before exploring further.

The UOSE system is an object-semantic execution platform for enterprise data, knowledge, and business systems. It converts models, tables, fields, metrics, entities, services, APIs, and operations from external resources into a unified ontology object space that can be queried, governed, and executed. Agents no longer need to deal directly with scattered APIs, databases, or documents; instead, they complete discovery, planning, simulation, execution, and auditing within one semantic contract. Use the system: Open UOSE

​
Core Problem

Enterprise data and systems are often fragmented in four ways:
  • Resource fragmentation: BI semantic models, SAP OData, knowledge bases, databases, and business APIs all use different protocols.
  • Semantic fragmentation: the same business object may have different names, fields, granularity, and constraints in different systems.
  • Execution fragmentation: querying, writing, approvals, replay, and auditing are distributed across different flows.
  • Agent fragmentation: agents can easily start composing SQL, API parameters, or business meanings from natural language guesses.
The product goal of the UOSE system is to converge these fragments into a unified object space: objects have types, attributes, relationships, constraints, actions, and evidence; actions have input contracts, risk levels, policy decisions, and audit results.

​
Unified Object-Semantic Execution

UOSE is the product implementation of Unified Object-Semantic Execution. It emphasizes three kinds of unification:
  • Object: convert identifiable objects in external resources into entity instances, such as semantic_cube, sap_odata_entity_set, knowledge_entity, and database_table.
  • Semantic: persist business relationships, field meanings, aliases, constraints, and context between objects as an ontology snapshot.
  • Execution: convert executable capabilities of objects into actions, such as semantic_model.query_cube_slice, sap_odata.read_collection, and database.query_select.
Therefore, the UOSE system is not simply a data catalog or a tool invocation gateway. Its core value is that, after a resource is connected, it automatically gains semantic visibility, action discoverability, controllable execution, and traceable results.

​
Product Loop

A typical loop includes:
  1. An administrator registers a resource and configures connection secrets and capabilities.
  2. The system syncs external resource metadata and generates canonical ontology IR.
  3. The ontology layer publishes a snapshot and projects it into entity, relation, and action instances.
  4. The Agent obtains the minimum necessary context through queryEntities, getEntityNeighborhood, and discoverActions.
  5. Before execution, the Agent calls simulateAction; after policy and parameter checks pass, it calls executeAction.
  6. The system records audit information; high-risk actions enter the approval queue according to policy.

​
Product Boundaries

The UOSE system is responsible for giving external resources semantics, executable actions, and governance controls, but it does not replace the source systems themselves:
  • It does not copy business ownership from source systems.
  • It does not bypass source system permissions, authentication, or concurrency control.
  • It does not require every resource to become a heavy ontology from the start.
  • It does not allow Agents to freely call arbitrary backend APIs.
It provides a unified control plane and execution plane for enterprise agents to access resources.