Data-Centric Application Platform

Create Data Centric Applications

  • Easy for anyone to build applications quickly
  • All application data is stored as RDF
  • Data is accessible over web protocols to humans, programs and AI agents
  • Access Control is a first-class concept, not an after-thought

Build your business on a Semantic Layer

The Semantic Layer between data, apps, and AI

  • Use GenAI to build applications that access their data via a Semantic Layer.
  • The Semantic Layer gives every application the same understanding of your business: what things mean, how they relate, who can access them, and what can be done with them.
  • Leverage the capabilities of AI agents by giving them equal access to your data as humans, via the Semantic Layer.

The Data-Centric Manifesto

We didn't just sign it.
We built the platform for it.

GraphCentric was designed from the ground up to embody the principles of the Data-Centric Manifesto.

  • Data is a key asset of any person, organization, and society.

    GraphCentric treats data as the durable center of the system — not a byproduct of applications.

  • Data is self-describing and does not rely on an application for interpretation and meaning.

    All data is stored as RDF with rich metadata and link relations so meaning is explicit and machine-readable.

  • Data is expressed in open, non-proprietary formats.

    We use RDF, HTML, JSON-LD, and standard web formats — never proprietary silos.

  • Access to and security of the data is a responsibility of the enterprise data layer or the personal data vault, and not managed by applications.

    Named graphs are the primary access control boundary. Applications request data, and request to update data, through the governed layer.

  • Applications are allowed to visit the data, perform their magic and express the results of their process back into the data layer.

    Applications read from and write back into the shared graph using well-defined updates, preserving truth and lineage.

The problem

Most software keeps recreating the same business facts.

Traditional applications start with screens and workflows, then hide their data behind private schemas, permissions tables, integration queues, and application-specific APIs.

Every new product recreates a partial version of the organisation. Change becomes expensive, integrations become brittle, and AI agents inherit stale context, ambiguous semantics, and permissions that were bolted on after the fact.

Mission

Separate data from applications without separating it from execution.

GraphCentric is for organisations that want many applications, automations, and agents to operate over the same live, governed body of business knowledge.

The platform keeps the durable model, access boundary, mutation logic, and web representations together, so new applications become purposeful surfaces over shared reality rather than another source of data fragmentation.

What is GraphCentric?

The runtime for data-centric applications.

Graph of record

Model business reality as addressable graph data, linked documents, object content, and operational metadata that outlive any single application surface.

Policy at the data boundary

Use named graphs as the primary visibility boundary so people, services, and agents only see the information they are allowed to read or mutate.

Resources over routes

Publish pages, update endpoints, API resources, negotiated representations, and reactive streams from the same resource model over ordinary HTTP.

Promise-based deployment

Describe the intended state of resources, links, updates, templates, and objects, then converge the running domain toward that state without destructive assumptions.

Agent-readable by default

AI agents should not have to scrape the human page.

Some agents prefer Markdown: it is compact, text-first, and easier to summarize, cite, transform, or feed into a planning loop than rendered HTML.

A GraphCentric server can link to that machine-friendly representation through ordinary web discovery. A HEAD request to the canonical resource advertises both HTML and Markdown alternates with typed Link headers.

In fact, the content on this web page (served by our GraphCentric server) is accessible as Markdown formatted text.

Discover the alternate representation

curl -I https://graphcentric.com/index

Response header

link: <https://graphcentric.com/index.md>; rel="alternate"; type="text/markdown"; title="Public Index Markdown"

Platform principles

Shared data becomes an application platform when it can be safely used.

One model, many surfaces

Applications become views, forms, updates, and representations over shared resources instead of private databases joined later by fragile synchronisation code.

Governed change

Typed update definitions, transaction graphs, and explicit resource promises make application behaviour inspectable, repeatable, and safer to evolve.

Discoverable semantics

Resources advertise links, metadata, formats, and state-shaping queries so humans, software, and agents can inspect the system they are using at runtime.

The philosophy

Data-centric does not mean abandoning the web. It means using it properly.

The web proved that systems can evolve independently and still interoperate through runtime introspection, standard methods, shared semantics, and discoverable metadata. GraphCentric applies that lesson inside the enterprise data layer, where most application platforms still rely on hidden contracts and duplicated state.

  • Independent evolution matters. At scale, systems cannot rely on tightly coupled, fixed contracts forever.
  • Runtime introspection matters. Agents need to discover capabilities, metadata, and representations as they operate.
  • HTTP still matters. Proper methods, headers, status codes, caching, and negotiation are part of the architecture, not legacy details.
  • Content negotiation matters. Machines and humans should be able to access the same resources through different representations.
  • Data-centric design matters. The center of gravity should move from application silos to durable, governed, interoperable information.

This sits naturally alongside the Data-Centric Manifesto.

Find out more

Get in touch

Want to explore what a semantic layer could do for your applications, data, or AI agents? Tell us what you are working on and we will follow up.

We couldn't submit your request. Please try again.

We will get back to you shortly.