Platform overview
Welcome to Loambase Foundry — an ontology-driven agent platform for developers and platform teams.
Its core goal: give enterprises one complete path “from extracting raw data sources” to “building top-level business applications”. Along the way, the platform introduces an Ontology semantic layer that turns cold physical database schemas into objects and relationships that business teams, developers — and even LLM agents — can understand directly.
What problem does it solve?
Traditional business system development typically faces these pain points:
- Data silos & duplicated work: every new app wires up the database again, rewrites its SQL, and ends up with inconsistent business criteria.
- No business semantics: frontend apps and LLMs are faced with physical tables like
t_order_info— expensive to interpret. - The path from data to app is too long: after data engineers clean the data, a backend still has to write APIs, a frontend still has to build pages — slow time to launch.
Loambase Foundry’s answer: one foundation, layered building.
The core usage path
On the platform, all data and business logic flow along this standard 4-step path:
1. Connect data
The platform offers many datasource connectors (MySQL, PostgreSQL, APIs, Kafka, and more). You sync physical tables from external systems into the platform as unified raw datasets.
2. Build pipelines
Raw data typically carries dirty data or awkward structures. With the visual pipeline builder you arrange filtering, cleansing, joins and aggregations like drawing a flowchart (DAG orchestration), producing high-quality derived datasets.
3. Model the ontology
This is the platform’s most important step. No SQL needed anymore: map high-quality datasets into business objects (e.g. Employee, Order), properties (e.g. employee name) and links (e.g. which department an employee belongs to). Together these form the enterprise-wide business semantic layer.
4. Ship apps & open access
Once the ontology is modeled, business capabilities are immediately usable:
- Business teams use the built-in App Builder: drag-and-drop components bound directly to ontology data, shipping business portals or analytics dashboards with zero code.
- Developers read objects and trigger actions through the strongly-typed SDK (React, Node.js, etc.) or the GraphQL API.
- AI & agents: the ontology model is naturally a structured toolset (Tools) — LLMs work directly on the semantic layer, eliminating the hallucination and privilege-escalation risks of letting AI hand-write SQL against physical tables.
Next step
Now that you know what problem the platform solves, continue with Core concepts to learn the platform vocabulary, or jump straight into the Quick start to ship your first app.